491+ terms every tech contractor and business owner should know. AI, web development, SEO, mobile, cloud, and business terminology.
Showing two different versions of a page to different visitors to see which performs better. Half see version A, half see version B, and data decides the winner. It's the scientific method applied to web design. Stop guessing, start testing.
Showing two versions of something to different audiences to see which performs better. Version A might have a blue button, Version B a green one. You split traffic 50/50 and measure which version gets more clicks. The winner becomes the new default. This removes guesswork from marketing decisions and lets data drive improvements. Even small changes can produce significant results when tested properly.
An automated email sent to shoppers who added items to their cart but didn't complete the purchase. These emails typically include a reminder of what was left behind, product images, and a link back to complete checkout. Some include a small discount to incentivize completion. Abandoned cart emails recover an average of 5 to 10% of abandoned carts, making them one of the highest-ROI email automations for e-commerce.
The portion of a web page visible without scrolling. Borrowed from newspaper terminology where the most important stories go above the physical fold of the paper. Your headline, value proposition, and primary CTA should all be above the fold. If a visitor has to scroll to understand what you do or how to take action, you've already lost a percentage of them.
Making websites usable for everyone, including people with disabilities. Screen readers, keyboard navigation, color contrast—it all matters.
Designing interfaces that adjust to different screen sizes, orientations, and device types, like foldable phones and tablets. Unlike responsive design which just reflows content, adaptive layouts may show completely different UI patterns for different screens.
A performance-based arrangement where you pay partners a commission for driving sales to your business. Someone promotes your product on their blog, social media, or email list. When someone buys through their unique link, they earn a percentage. You only pay for actual results, making it low-risk. Amazon runs the largest affiliate program in the world. Plenty of smaller businesses use affiliates too.
AI systems designed to operate independently, making decisions, using tools, and completing multi-step tasks without constant human guidance. Unlike chatbots that just respond, agentic AI plans and executes. Think autonomous assistant, not just smart autocomplete.
When AI doesn't just answer questions but actually takes actions on your behalf. Think: an AI that books your flights, sends emails, and updates your calendar without you clicking anything.
A development approach focused on shipping small, frequent updates and adapting to feedback. The opposite of spending months building something nobody wants.
An AI that can actually do things, not just chat. It can browse the web, run code, call APIs, and complete multi-step tasks on its own. Think AI with hands, not just a mouth.
Making sure AI systems actually do what humans want them to do. It's the problem of ensuring AI goals match our goals, especially as these systems get more powerful. If your GPS always routes you through toll roads, that's a tiny alignment problem.
A service that lets you send requests to an AI model and get responses back, usually charged per use. You don't run the model yourself, you just call their servers.
A single entry point that routes requests to different AI providers. It handles load balancing, fallbacks, caching, and logging so you don't have to.
Techniques for catching when AI makes stuff up. Methods include cross-referencing outputs against known facts, using multiple models to verify claims, and training classifiers to spot confident-sounding nonsense. Essential for any production AI system.
Coordinating multiple AI models or agents to work together on complex tasks. It's like being a conductor - you're directing which AI does what and when.
The practice of building AI systems that behave as intended without causing harm. Covers everything from preventing bias to making sure AI doesn't leak sensitive data.
A text description of an image that helps search engines understand what it shows. It also makes your site accessible to people using screen readers.
Accelerated Mobile Pages—Google's framework for making stripped-down, lightning-fast mobile pages. It was huge for a while, but most sites don't need it anymore since Core Web Vitals became the focus.
The clickable text in a hyperlink. "Click here" is bad anchor text. "Best running shoes for beginners" is good anchor text because it tells both users and search engines what the linked page is about. It's a ranking signal Google pays attention to.
Google's official IDE for Android development, built on JetBrains' IntelliJ platform. It includes an emulator, layout editor, profiling tools, and built-in support for Kotlin and Jetpack Compose. The standard tool for any serious Android work.
The average annual revenue per contract. If one customer pays $120K/year and another pays $60K/year, your average ACV is $90K. Enterprise SaaS companies obsess over ACV because it determines whether they need a sales team or can rely on self-serve.
A set of rules that lets different software talk to each other. It's how your app asks another service for data or tells it to do something.
A unique identifier used to authenticate requests to an API. It's like a password for your application. Never expose API keys in client-side code or public repositories. Rotate them regularly, and use different keys for different environments.
A publishing format (used by Google Play and now Apple) that lets the store generate optimized APKs for each device. Users download only what they need.
A small part of an iOS app that users can access instantly without downloading the full app. Scan a QR code at a restaurant and the ordering interface appears immediately, no App Store visit required. Limited to 15MB.
Android's equivalent of Apple's Universal Links. They let you claim ownership of web URLs so they open in your app automatically. When a user taps a link to your domain, Android opens your app instead of the browser, no disambiguation dialog needed.
A security process that cryptographically proves your app comes from you and hasn't been tampered with. Both Apple and Google require it.
SEO for app stores. Optimizing your app's title, description, screenshots, and reviews to rank higher in App Store and Google Play searches.
MRR multiplied by 12. If your MRR is $50,000, your ARR is $600,000. It's the standard way SaaS companies describe their size. "$1M ARR" is a common milestone that means you've found real traction. Enterprise SaaS companies think in ARR, consumer ones in MRR.
A web framework built for content-heavy sites like blogs, marketing pages, and documentation. It ships zero JavaScript by default and lets you use components from React, Vue, or Svelte in the same project. Pages are lightning fast because they're mostly static HTML.
The core idea behind transformers that lets AI focus on the most relevant parts of the input when generating output. It's like how you skim a page and your eyes jump to the important bits. This is what makes modern language models actually understand context instead of just reading left to right.
Rules for deciding which marketing touchpoint gets credit for a conversion. Did the Google ad, email, or social post drive the sale? Different models give different answers. Last-click gives all credit to the final touch, while multi-touch spreads it around.
A rule for deciding which marketing touchpoint gets credit for a conversion. Did the customer convert because of the Google ad they clicked first, the email they opened second, or the retargeting ad they clicked last? First-click attribution credits the first touch. Last-click credits the last. Data-driven attribution distributes credit across all touchpoints based on their contribution. Getting attribution right is crucial for smart budget allocation.
Proving who you are. It's the login process where you show your credentials like email and password to verify your identity.
Deciding what you can do after you've logged in. Authentication asks who are you, authorization asks what are you allowed to access.
Automatically adding or removing servers based on demand. You handle traffic spikes without overpaying during quiet periods.
A car or truck that drives itself using AI, sensors, cameras, and lidar. Levels range from 1 (cruise control) to 5 (no steering wheel needed). Most "self-driving" cars today are level 2-3, meaning they still need a human paying attention.
The average amount a customer spends per order. Calculate it by dividing total revenue by number of orders. If you made $10,000 from 200 orders, your AOV is $50. Increasing AOV is often easier than increasing the number of orders. Strategies include upselling (suggesting a premium version), cross-selling (recommending related products), free shipping thresholds, and bundle deals.
The largest cloud computing platform, offering over 200 services from virtual servers to machine learning. About a third of all cloud infrastructure runs on AWS. It's incredibly powerful but notorious for its confusing pricing and overwhelming number of services.
The server-side code that handles data, business logic, and database operations. Users never see it directly, but it powers everything behind the scenes.
Links from other websites pointing to yours. Google sees these as votes of confidence—the more quality sites linking to you, the more trustworthy you appear to search engines.
A prioritized list of everything that needs to be built. The top items get worked on next. Things at the bottom might never get done—and that's okay.
Sending a bunch of AI requests at once and getting all the results back together, instead of one at a time. It's usually cheaper but slower since you're waiting for the whole batch.
A standardized test to measure how well an AI model performs. It's how we compare apples to apples when evaluating different models.
Using fingerprints, face recognition, or iris scanning to verify identity on mobile devices. Face ID and Touch ID on iPhones, fingerprint sensors on Android. More secure and convenient than passwords for most mobile users.
A distributed ledger where transactions are grouped into blocks and chained together cryptographically. Every participant has a copy, and changing historical records is practically impossible. It's the technology underlying Bitcoin, Ethereum, and all cryptocurrencies.
Running two identical production environments where one serves live traffic while the other gets updates. When the new version is ready, you flip traffic instantly—zero downtime.
Building a business with your own money and revenue instead of taking outside investment. You grow slower but keep 100% ownership and full control. Basecamp, Mailchimp (before acquisition), and Craigslist were famously bootstrapped.
The percentage of visitors who leave your site after viewing only one page. In GA4, it's been replaced by "engagement rate" (the inverse). A high bounce rate might mean your content didn't match what users expected, or it perfectly answered their question.
The percentage of visitors who leave your site after viewing only one page. If 100 people visit and 60 leave without clicking anything else, your bounce rate is 60%. High bounce rates on landing pages usually signal a problem: slow load time, confusing design, or a mismatch between what brought the visitor and what the page delivers. For blog posts, high bounce rates are more normal since readers often find their answer and leave.
How familiar people are with your business. When someone needs a plumber and your name is the first one that pops into their head, that's brand awareness working. It's hard to measure directly but shows up in branded search volume (people Googling your name), direct traffic, and word-of-mouth referrals. Building brand awareness is a long-term investment that pays off when people are ready to buy.
Built-in features that browsers provide to JavaScript, like accessing the camera, reading device sensors, storing data locally, or sending notifications. They're how web apps get capabilities that used to require native applications.
An all-in-one JavaScript runtime, bundler, and package manager that's significantly faster than Node.js. It runs TypeScript natively without compilation, installs npm packages in seconds, and aims to be a drop-in Node.js replacement. Written in Zig.
The total size of your JavaScript after it's been packaged for production. Smaller bundles mean faster load times for users.
How fast a startup spends its cash reserves, usually measured monthly. If you have $1M in the bank and spend $100K/month, your burn rate is $100K and you have 10 months of runway. Every founder watches this number obsessively.
How much you spend to get one new customer, including ads, sales, and marketing costs. If you spend $10,000 on marketing and get 100 customers, your CAC is $100. The goal is to keep CAC well below what a customer is worth over their lifetime.
How many months it takes to earn back what you spent acquiring a customer. If your CAC is $300 and customers pay $100/month with 70% margin, payback is about 4.3 months. Under 12 months is good. Over 18 months is dangerous for most startups.
Storing frequently used data closer to where it's needed so you don't have to fetch it every time. Makes everything faster.
The specific action you want someone to take on your page. "Get a Free Quote," "Book a Consultation," "Start Your Trial." A good CTA is specific about what happens next, uses action words, and stands out visually. Bad CTAs are vague ("Submit," "Click Here") or invisible. Every page on your website should have one primary CTA that's impossible to miss.
Rolling out changes to a small percentage of users first before going fully live. If something breaks, only a few people are affected instead of everyone.
The "official" version of a page when you have duplicate or similar content at multiple URLs. It tells Google which one to index and rank, preventing duplicate content issues.
A cross-platform runtime by the Ionic team that turns web apps into native mobile apps. You build with standard web tech (HTML, CSS, JS), and Capacitor wraps it in a native shell with access to device APIs like camera, GPS, and push notifications.
A network of servers around the world that stores copies of your files. Users download from the nearest server instead of your origin.
Getting an AI to reason step-by-step through a problem instead of jumping straight to an answer. It dramatically improves accuracy on complex tasks like math and logic.
A formal ask to modify something after we've already agreed on the scope. Not a bad thing - projects evolve. But it usually affects timeline and budget.
Intentionally breaking things in production to test how resilient your system is. Netflix pioneered this with their "Chaos Monkey" that randomly kills servers. If your system can survive deliberate sabotage, it'll handle accidental failures too.
An AI program that talks with users through text or voice. Modern chatbots use LLMs to have actual conversations instead of just matching keywords to canned responses.
Reducing friction in the buying process to prevent cart abandonment. Best practices include guest checkout (don't force account creation), minimal form fields, multiple payment options, progress indicators, security badges, and transparent pricing (no surprise fees at the last step). Every additional step in checkout loses customers. The best checkouts are one page with as few fields as possible.
The percentage of customers who cancel their subscription in a given period. 5% monthly churn means you lose half your customers every year. Even a small improvement in churn can dramatically change your business trajectory. Retention is cheaper than acquisition.
The percentage of customers who stop doing business with you over a given period. If you start the month with 100 customers and lose 5, your monthly churn rate is 5%. High churn means you're constantly replacing lost customers just to stay flat. Reducing churn by even 1-2% can dramatically increase revenue because keeping existing customers is way cheaper than acquiring new ones.
Short for Continuous Integration/Continuous Deployment. It automates testing and deploying your code so you can ship faster with fewer bugs.
The percentage of people who click after seeing your link, ad, or email. In search, it's impressions divided by clicks. In email marketing, it's how many recipients clicked a link. A higher CTR means your messaging resonates with your audience.
The percentage of people who click on your link, ad, or email after seeing it. If 100 people see your ad and 3 click it, your CTR is 3%. Higher CTR means your message resonates with your audience. Low CTR means something needs to change: the headline, the image, the targeting, or the offer itself. Average CTRs vary wildly by platform and industry.
Renting computing resources like servers, storage, and databases from a provider instead of owning them. You pay for what you use, scale up or down instantly, and let someone else worry about hardware failures. AWS, Google Cloud, and Azure dominate this space.
Measures how much stuff jumps around while your page loads. You know when you're about to tap a button and an ad loads, pushing everything down? That's layout shift.
Breaking your JavaScript into smaller chunks that load on demand. Users only download the code they actually need for the page they're on.
Grouping customers by when they signed up and tracking their behavior over time. January's cohort might retain 60% after 6 months while March's retains 75%. This tells you whether product improvements are actually working, not just whether you're growing.
Grouping users by a shared characteristic (like signup date) and tracking their behavior over time. Instead of looking at all users as one blob, you see how January signups behave versus February signups. This reveals trends like whether newer users are more or less engaged than older ones. Especially useful for subscription businesses tracking retention and predicting churn.
A recommendation technique that finds users similar to you and suggests things they liked. If people who watched the same 10 movies as you also loved movie #11, you'll probably like it too. It's the "people who bought this also bought" engine.
A reusable piece of UI—like a button, card, or navigation bar. You build it once and use it everywhere, keeping your code consistent and maintainable.
AI that can see and understand images and video. It lets machines identify objects, read text from photos, detect faces, and understand what's happening in a scene.
A lightweight, isolated package containing your app and its dependencies. Like a VM but way faster and smaller because it shares the host OS.
A storage service for Docker container images. You push built images there and pull them when deploying. Docker Hub is the public default, but most companies use private registries like AWS ECR or Google Artifact Registry for security.
An SEO strategy where you create a main "pillar" page on a broad topic, surrounded by detailed "cluster" pages on related subtopics, all linked together. It shows search engines you've covered a topic thoroughly and builds topical authority.
Attracting customers by creating useful content instead of directly advertising to them. Blog posts, videos, podcasts, guides, infographics. The idea is that if you consistently help people with their problems for free, they'll trust you enough to pay you when they need more help. It's a long game that compounds over time, unlike ads that stop working when you stop paying.
HTTP headers that tell browsers which resources your page is allowed to load. It prevents XSS attacks by blocking unauthorized scripts from running. If a script isn't on your approved list, the browser won't execute it, even if an attacker injected it.
How much text an AI can "see" at once—its working memory. Bigger windows mean you can feed it longer documents, but they cost more and can slow things down.
Automatically deploying code to production after it passes all tests. No manual "deploy" button—if tests pass, it ships.
The practice of automatically building and testing code every time someone pushes changes. Catches problems early instead of at release time.
The specific sequence of steps a visitor takes to become a customer on your website. Views ad, clicks through, lands on page, reads content, fills out form, becomes lead, gets follow-up, becomes customer. Analyzing each step shows you where people drop off and where to focus optimization efforts. Even small improvements at the top of the funnel compound into big results at the bottom.
The percentage of visitors who complete your desired action. If 100 people visit and 3 buy, that's a 3% conversion rate. A small improvement here can dramatically impact revenue. Going from 2% to 3% is a 50% increase in sales.
The percentage of visitors who complete a desired action. If 1,000 people visit your page and 30 fill out your form, your conversion rate is 3%. This is arguably the most important metric for any business website because it directly ties traffic to business results. Average conversion rates hover around 2-3% for most industries, but well-optimized pages can hit 10% or higher.
Improving your online store to get more visitors to buy. Specifically for e-commerce, this means optimizing product pages, simplifying checkout, adding trust signals, improving site search, and enhancing mobile shopping experience. A 1% increase in conversion rate on a store doing $100k/month in revenue translates to $12,000+ in additional annual revenue without spending more on traffic.
Measuring when visitors complete a desired action on your site, like buying something, signing up, or downloading a file. Without conversion tracking, you're flying blind. You might be getting tons of traffic that never converts, and you'd never know.
The process of recording when a website visitor completes a valuable action. Form submissions, purchases, phone calls, chat initiations. Without conversion tracking, you know how many people visit but not how many become customers. Setting this up is the single most important analytics task for any business website. Everything else is just counting eyeballs.
Google's three key metrics for measuring user experience: loading speed (LCP), interactivity (FID), and visual stability (CLS). These directly affect your search rankings.
Three specific metrics Google uses to measure user experience: Largest Contentful Paint (loading speed), Interaction to Next Paint (interactivity), and Cumulative Layout Shift (visual stability). Google uses these as ranking factors. Sites that pass Core Web Vitals get a small ranking boost. Sites that fail may rank lower. Check yours in Google Search Console or PageSpeed Insights.
A security rule that controls which websites can make requests to your server. It prevents random sites from accessing your API on behalf of users.
How much you spend to get one new customer. Calculate it by dividing total marketing spend by the number of new customers gained. If you spent $1,000 on ads and got 10 customers, your CPA is $100. This is the most important marketing metric because it tells you whether your marketing is actually profitable. If your CPA is higher than what the average customer is worth, you're losing money.
How much you pay each time someone clicks your ad. In Google Ads, you bid on keywords and pay when someone clicks. CPCs range from $0.50 for low-competition keywords to $50+ for competitive industries like legal or insurance. CPC alone doesn't tell you much. A $10 click that turns into a $5,000 customer is a great deal. A $0.50 click that bounces immediately is wasted money.
What you pay for each chunk of text the AI processes. Tokens are roughly 4 characters or 3/4 of a word. Input and output tokens usually have different prices.
The cost of 1,000 ad impressions. If you pay $5 CPM, you're paying $5 for your ad to be shown 1,000 times. This pricing model is common on social media and display advertising. CPM-based campaigns are better for brand awareness than direct response. If you need clicks and conversions, CPC bidding usually makes more sense than CPM.
Automatically collecting and reporting app crashes with stack traces, device info, and reproduction steps. Sentry and Firebase Crashlytics catch crashes in real-time so you can fix them before negative reviews pile up in the app store.
The number of pages Google will crawl on your site within a given timeframe. Big sites need to worry about this—if Google can't crawl your important pages, they won't get indexed.
Building one app that runs on both iOS and Android. Saves time and money compared to building two separate native apps.
Recommending related products alongside what a customer is buying. "People who bought this also bought..." or "Complete the look with these accessories." Amazon generates about 35% of its revenue from cross-selling recommendations. For smaller stores, displaying related products on product pages and in the cart is the simplest implementation. Keep recommendations relevant and limited to 3-4 items.
The browser downloads a minimal HTML shell, then JavaScript builds the page. Fast subsequent navigations, but users might see a loading spinner on first visit.
An attack that tricks your browser into making unwanted requests to a site where you're logged in. You visit a malicious page, and it secretly submits a form to your bank's site using your active session. CSRF tokens in forms prevent this.
Writing CSS styles directly in your JavaScript files instead of separate .css files. Libraries like styled-components and Emotion make this work. It keeps styles scoped to components and lets you use JavaScript logic in your styling.
The strategies and costs involved in getting new customers to your online store. For e-commerce, the main channels are paid ads (Google Shopping, Meta), SEO and content marketing, email marketing, social media, and influencer partnerships. The key metric is customer acquisition cost (CAC) versus average order value and lifetime value. Sustainable e-commerce businesses acquire customers profitably from multiple channels.
The total cost of acquiring a new customer, including all marketing and sales expenses divided by the number of new customers. If you spend $10,000 on marketing and sales in a month and gain 50 new customers, your CAC is $200. Compare CAC to Customer Lifetime Value. If CLV is significantly higher than CAC, your business model works. If they're close or CAC is higher, you need to either reduce acquisition costs or increase customer value.
The complete experience a customer has with your business, from first discovering you to making a purchase and beyond. Mapping this journey reveals touchpoints where customers interact with your brand, friction points where they get stuck, and opportunities to improve the experience. Most businesses only think about the purchase moment and ignore everything before and after it.
The total revenue you can expect from a single customer over their entire relationship with your business. If a customer spends $50/month and stays for 2 years, their CLV is $1,200. This number determines how much you can afford to spend on acquiring a customer. A business with $1,200 CLV can profitably spend $200 to get a customer. A business with $50 CLV can't.
Product ratings and written feedback from buyers displayed on product pages. Reviews dramatically impact purchase decisions. Products with reviews convert at nearly 4x the rate of products without them. Even negative reviews help because they build trust that the positive ones are real. Actively collect reviews through post-purchase emails and make it easy for customers to leave them.
A public catalog of known security vulnerabilities, each with a unique ID. When a bug is found in software, it gets a CVE number like CVE-2024-1234 so everyone's talking about the same issue. Checking CVEs for your dependencies is basic security hygiene.
Applications that run on a blockchain network instead of centralized servers. No single company controls them, and they can't be shut down by any one authority. Uniswap (decentralized exchange) and OpenSea (NFT marketplace) are well-known examples.
A visual display of your most important metrics in one place. Instead of digging through multiple reports, a dashboard puts your key numbers front and center. Good dashboards show 5 to 8 metrics that actually drive decisions. Bad dashboards try to show everything and end up showing nothing useful. Tools like Google Looker Studio (free), Databox, and Klipfolio create automated dashboards.
The process of tagging data with the correct answers so AI can learn from it. Someone has to look at 100,000 photos and mark which ones contain cats before the model can learn to find cats. It's often the most expensive and time-consuming part of building AI.
Making business choices based on data rather than gut feeling. Instead of redesigning your homepage because you're bored of it, you redesign it because data shows a 70% bounce rate and 0.5% conversion rate. Sounds obvious, but most businesses still make marketing decisions based on opinions and preferences rather than measurable evidence. The tools are free. The data is there. Most people just don't look at it.
Where your app stores all its data—user accounts, posts, orders, everything. Think of it as a really organized spreadsheet that code can read and write to.
An attack that floods your server with so much traffic it can't serve real users. Thousands of compromised machines (a botnet) all hit your site simultaneously. It's like a flash mob blocking a store entrance so customers can't get in.
Machine learning using neural networks with many layers (hence "deep"). More layers can learn more complex patterns. This is what powers image recognition, language models, and most modern AI.
Links that open a specific screen inside your app instead of just the home screen. Essential for marketing campaigns and user experience.
AI-generated fake media, usually video or audio, that makes it look like someone said or did something they didn't. The technology can swap faces in videos, clone voices from a few seconds of audio, and create entirely fictional scenarios that look disturbingly real.
A tangible thing we hand over to you - could be code, designs, documentation, or a working feature. Something you can actually see and verify.
A JavaScript runtime created by the same person who made Node.js, built to fix Node's design mistakes. It supports TypeScript natively, has built-in security permissions, and uses URL-based imports instead of node_modules. Growing steadily but Node.js still dominates.
A culture and set of practices that bridges development and operations. The goal is faster, more reliable releases through automation and collaboration.
The AI architecture behind image generators like Stable Diffusion and DALL-E. It works by learning to remove noise from images, then generating new images by starting with pure noise and gradually cleaning it up into something meaningful.
Your plan for getting back online when things go catastrophically wrong. Covers backups, failover systems, and how fast you can recover.
Telling Google to ignore specific backlinks pointing to your site. If spammy or toxic sites link to you and you can't get them removed, you submit a disavow file through Google Search Console. Use it carefully. Most sites don't need it.
The research and planning period before actual development starts. This is where we figure out what you actually need, not just what you think you need.
Tracking a single request as it flows through multiple microservices. When a user hits your app, the request might touch 10 different services. Distributed tracing connects all those hops so you can see exactly where things slow down or fail.
The internet's phone book. It translates human-friendly domain names into the IP addresses computers actually use.
The most popular tool for creating containers. It packages your app with everything it needs to run, so it works the same everywhere.
Your website's address on the internet. It's the name people type to find you, like example.com.
A score (1-100) that predicts how well your entire website will rank in search results. It's based on factors like backlinks, age, and overall trustworthiness. Higher is better.
Ahrefs' proprietary score (0-100) measuring the strength of a website's backlink profile. It's similar to Moz's Domain Authority but uses a different calculation. Higher ratings generally correlate with better ability to rank for competitive keywords.
A simpler alternative to RLHF for aligning AI models with human preferences. Instead of training a separate reward model, it directly optimizes the language model using pairs of preferred and rejected responses. Faster to implement and increasingly popular.
A series of automated emails sent over time to nurture leads. Someone downloads your guide, then gets Email 1 the next day, Email 2 three days later, and so on. Each email builds on the last, gradually moving the recipient toward a purchase decision. Well-designed drip campaigns can run on autopilot and convert leads while you sleep.
A lightweight TypeScript ORM that stays close to SQL instead of abstracting it away. Developers who like writing SQL but want type safety gravitate toward Drizzle. It's thinner than Prisma, meaning less magic and more control over your queries.
An e-commerce model where you sell products you don't physically stock. When a customer orders, you forward the order to a supplier who ships directly to the customer. You never handle inventory. The appeal is low startup cost and zero inventory risk. The downsides are thin margins, no control over shipping speed or product quality, and intense competition since the barrier to entry is so low.
How long someone stays on your page after clicking from search results before going back to Google. Longer dwell time suggests your content was useful. If people click back in 5 seconds, Google notices and might rank you lower.
Google's quality guidelines: Experience, Expertise, Authoritativeness, and Trustworthiness. It's how Google decides if your content is actually worth ranking—do you know what you're talking about, and can users trust you?
Processing data closer to where users are instead of sending everything to a central server. Your code runs at data centers scattered around the world, so a user in Tokyo gets served by a nearby server, not one in Virginia. Reduces latency dramatically.
Code that runs at CDN locations close to your users instead of one central server. Lower latency because the response doesn't have to travel across the world.
Using email to communicate with potential and existing customers. Newsletters, promotional offers, product updates, drip campaigns. It consistently delivers the highest ROI of any marketing channel at roughly $36 returned for every $1 spent. The key advantage is ownership: your email list is yours. Social media platforms can change algorithms or ban accounts. Your email list is a direct line to your audience.
A way to turn words, sentences, or documents into numbers that capture their meaning. Similar concepts get similar numbers, which lets AI find related content even if the exact words don't match.
Scrambling data so only authorized parties can read it. Without the decryption key, the data looks like random gibberish. It protects everything from your WhatsApp messages to your credit card numbers during online purchases.
Encryption where only the sender and receiver can read the messages. Not even the service provider can decrypt them. WhatsApp and Signal use E2EE, meaning even if their servers get hacked, your messages stay private.
In GA4, the percentage of sessions that qualify as "engaged." A session counts as engaged if it lasts longer than 10 seconds, has a conversion event, or includes 2+ pageviews. It's the opposite of bounce rate and gives a more nuanced view of how visitors interact with your site. An engagement rate above 60% is generally healthy. Below 40% suggests your content or user experience needs improvement.
Recording specific interactions visitors have with your website beyond pageviews. Button clicks, form submissions, video plays, file downloads, scroll milestones. GA4 is entirely event-based, meaning every interaction is tracked as an event. Setting up custom events for actions that matter to your business gives you much deeper insight into how people use your site.
A design pattern where services react to events instead of directly calling each other. When a user signs up, an "UserCreated" event fires, and the email service, analytics service, and CRM each independently respond to it. Loose coupling at its finest.
The percentage of views on a page that were the last in the session. Different from bounce rate because it includes people who visited other pages first. A high exit rate on your checkout page is bad (people are abandoning their purchase). A high exit rate on your thank-you page is normal (they completed the action and left). Context matters when interpreting exit rates.
Additional revenue from existing customers through upgrades, upsells, or add-ons. It's cheaper to grow revenue from happy customers than to find new ones. If your expansion revenue exceeds your churn, you have negative churn, the holy grail of SaaS.
AI systems that can explain why they made a specific decision. Most AI is a black box. If your loan gets denied by AI, explainable AI can tell you it was because of your credit utilization rate, not just "computer says no."
A toolkit that makes React Native development way easier. Handles the annoying native stuff so you can focus on building features instead of fighting Xcode.
A switch in your code that lets you turn features on or off without deploying new code. Super useful for testing features with specific users or killing a broken feature instantly.
The answer box at the top of Google that directly answers a question. It's position zero—above all other results. Landing here can massively boost your traffic.
Training AI across multiple devices or organizations without sharing the actual data. Each participant trains locally and only sends model updates, not raw data. It's how your phone can improve autocorrect without sending your messages to Apple or Google.
Giving an AI a few examples of what you want before asking it to do the task. Those examples teach it the pattern without needing to retrain the whole model.
The time between when a user first clicks something and when the browser actually responds. If your site feels frozen when people try to interact, your FID is bad.
Teaching an existing AI model new tricks by training it on your specific data. It's like hiring someone with general skills, then training them on how your company does things.
A security system that monitors and controls network traffic based on rules you set. It's the bouncer at the door, deciding which connections get in and which get blocked. Can be hardware, software, or cloud-based.
A contract where the total cost is agreed upfront for a defined scope. You know exactly what you'll pay, but changes outside the original scope cost extra.
Google's cross-platform framework using the Dart language. Known for smooth animations and a single codebase that works on mobile, web, and desktop.
A massive AI model trained on huge datasets that can be adapted for many different tasks. Think of it as the base layer that other applications build on top of.
A pricing model where the basic product is free and you charge for premium features. Spotify, Dropbox, and Slack all use freemium. The free tier gets users in the door, and the paid tier captures value. Conversion rates from free to paid typically range 2-5%.
The part of a website users actually see and interact with—buttons, forms, animations, everything in the browser. It's built with HTML, CSS, and JavaScript.
The entire process from receiving an order to delivering it to the customer. Picking, packing, shipping, and tracking. You can handle fulfillment yourself (self-fulfillment), use a third-party logistics provider (3PL), or use a platform like Amazon FBA. The right choice depends on your volume, product type, and budget. Fast, reliable fulfillment is a competitive advantage because customers expect quick delivery.
A developer who works on both frontend and backend. They can build an entire application from the database to the user interface.
When an AI can decide to call specific functions or tools you've defined, returning structured data instead of just text. It's how you connect AI to real actions and data.
Raising money from investors to fund your startup. Rounds go from pre-seed ($100K-$1M) through Series A, B, C and beyond. You give up equity in exchange for capital. Each round gets bigger but you own less of the company.
A chart showing how many users progress through each step of a process on your website. If 1,000 people view your product page, 500 click "add to cart," 200 start checkout, and 100 complete the purchase, the funnel visualization shows exactly where people drop off. This identifies the biggest leaks in your conversion process so you can focus optimization where it matters most.
Two neural networks competing against each other. One generates fake content, the other tries to detect it's fake. They keep pushing each other to improve until the generator creates incredibly realistic output. GANs pioneered AI image generation before diffusion models took over.
AI that creates new content like text, images, music, code, or video. Unlike traditional AI that classifies or predicts, generative AI produces original output. ChatGPT writing essays, Midjourney creating art, and Suno making music are all generative AI.
Touch-based interactions like swipe, pinch, long press, and pull-to-refresh that feel natural on mobile devices. Well-implemented gestures make apps feel fluid and intuitive. Badly implemented ones frustrate users who accidentally trigger them.
A popular file format for storing quantized AI models that can run on regular consumer hardware. If you want to run an LLM on your laptop without a fancy GPU, you're probably downloading a GGUF file. It replaced the older GGML format.
Using Git as the single source of truth for infrastructure and deployment. Every change goes through a pull request, and automated tools sync the Git state to your actual servers. If it's not in Git, it doesn't exist. Rollbacks are just git reverts.
The moment your software becomes available to real users. Launch day. Everything before this was prep - now it's real.
A specific user action you track as valuable in Google Analytics. Form submissions, phone calls, purchases, video plays, file downloads. In GA4, these are called "conversion events." Setting these up is critical because without them, analytics just tells you how many people visited. With conversions configured, it tells you how many people did something that matters to your business.
Google's current analytics platform that tracks how users interact with your website or app. It replaced Universal Analytics in 2023 with an event-based model. Everything is an "event" now, from page views to purchases, which makes tracking more flexible but the learning curve is steep.
The current version of Google's free web analytics platform. It replaced Universal Analytics in 2023 with an event-based tracking model. GA4 tracks users across devices, integrates with Google Ads, and uses machine learning to fill in data gaps from privacy restrictions. The learning curve is steeper than the old version, but it's more powerful and better suited to modern privacy-first web browsing.
Google's free tool that shows how your site performs in search. It tells you which queries bring traffic, which pages are indexed, and what technical problems Google found. If you only use one SEO tool, this should be it.
A free Google tool that shows how your site appears in search results. Which queries bring up your pages, your average ranking position, click-through rates, and technical issues Google found while crawling. If Google Analytics tells you what happens on your site, Search Console tells you what happens before people get there. Every website owner should have this set up and check it monthly at minimum.
Google's product search and advertising platform. When someone searches for a product, Shopping ads show product images, prices, and store names directly in search results. You need a Google Merchant Center account and a product feed to use it. Shopping ads often convert better than text ads for products because customers see the price and image before clicking, which pre-qualifies them.
Generative Pre-trained Transformer—OpenAI's family of language models. "Generative" means it creates content, "Pre-trained" means it learned from existing data, "Transformer" is the architecture.
An alternative to REST that lets you ask for exactly the data you need in a single request. No more getting too much or too little data.
Safety measures that keep AI from going off the rails - blocking harmful outputs, enforcing topic boundaries, and catching prompt injection attempts before they cause problems.
Unconventional marketing tactics that get attention without big budgets. Flash mobs, street art, viral stunts, creative giveaways. The idea is to create memorable experiences that generate word-of-mouth and social sharing. It works best for local businesses and startups that can't compete on ad spend but can compete on creativity. High risk, high reward. When it works, the ROI is enormous.
When AI confidently makes stuff up. It'll invent facts, cite fake sources, or describe things that don't exist—all while sounding completely sure of itself. This is why you verify AI output.
An e-commerce architecture where the front-end (what customers see) is completely separate from the back-end (product management, checkout, inventory). This lets you build any front-end experience you want without being limited by your e-commerce platform's templates. Shopify Hydrogen, BigCommerce with Next.js, and Medusa are popular headless options. More flexible but requires developer resources.
A visual representation of where visitors click, move their mouse, and scroll on your page. Hot spots (usually red) show heavy activity. Cold spots (blue) show areas nobody interacts with. Heatmaps reveal whether people actually see your CTA, how far they scroll, and which elements attract attention. Tools like Hotjar and Microsoft Clarity generate heatmaps. They often expose surprising user behavior.
Visual overlays showing where users click, scroll, and move their cursor on your pages. Hot spots (red/orange) show high activity, cold spots (blue) show ignored areas. They reveal what users actually do versus what you think they do.
Pre-packaged Kubernetes application definitions that make deploying complex apps as easy as running one command. Instead of writing dozens of Kubernetes config files, you install a Helm chart with your settings and it handles the rest.
React functions that let you use state and other features in functional components. useState, useEffect, useContext—they all start with "use."
Adding more servers to handle increased load instead of making existing servers bigger. It's like opening more checkout lanes instead of making one cashier work faster. Most cloud architectures prefer horizontal scaling because it's practically limitless.
The service that stores your website files and makes them accessible on the internet. No hosting means nobody can visit your site.
A development feature that lets you see code changes instantly without restarting the app. Saves tons of time during UI work.
HTML tags that tell search engines which language and regional version of a page to show to different users. If you have English, Spanish, and French versions of your site, hreflang ensures Google shows the right one based on the searcher's location and language.
HTTP with encryption. The S means your connection is secure and nobody can eavesdrop on what you're sending or receiving.
An app that's basically a website wrapped in a native container. One codebase works on both platforms, but performance isn't as good as native.
When server-rendered HTML gets "activated" by JavaScript in the browser. The page looks ready instantly, then becomes interactive once the JS loads and attaches event listeners.
AI that looks at an image and assigns it to a category. Is this a cat or a dog? Is this X-ray normal or abnormal? It's one of the oldest and most reliable computer vision tasks, and it powers everything from photo organization to medical diagnosis.
Instead of updating servers in place, you build entirely new ones with the changes baked in and replace the old ones. Servers are treated like cattle, not pets. No one SSHs in to fix things. This eliminates configuration drift and "works on my machine" problems.
One instance of your content being displayed to someone. If your ad shows up in 1,000 feeds, that's 1,000 impressions. Impressions don't mean people saw or engaged with your content. They just mean it appeared on screen. A useful metric for awareness campaigns, but don't confuse impressions with actual engagement. Millions of impressions with zero clicks means your content isn't resonating.
Buying stuff inside an app—subscriptions, extra features, virtual goods. Apple and Google take a 15-30% cut of every transaction.
When Google adds your page to its database so it can show up in search results. If a page isn't indexed, it basically doesn't exist as far as Google is concerned.
When AI actually runs and generates output. Training teaches the model; inference is when it does its job. This is what you're paying for with API calls.
Partnering with people who have engaged audiences to promote your product or service. Doesn't have to mean paying celebrities millions. Micro-influencers with 5,000 to 50,000 followers often produce better ROI because their audiences are more niche and engaged. The key is authenticity. Audiences can spot a forced promotion instantly, so work with influencers who genuinely align with your brand.
AI that pulls structured data from unstructured text. Give it a contract and it extracts dates, parties, dollar amounts, and key terms. Give it a resume and it pulls out skills, experience, and education. It turns messy documents into clean data.
Managing servers, databases, and networks through config files instead of clicking around dashboards. Version controlled, repeatable, and way less error-prone.
Android's equivalent of App Clips. Users can run parts of your app from a link without installing it. Great for try-before-you-install experiences and reducing friction for first-time users who don't want to commit to a full download.
Links between pages on your own website. They help search engines understand your site structure, distribute page authority, and guide users to related content. Most sites underdo internal linking. It's one of the easiest SEO wins you're probably ignoring.
Optimizing your website to rank in multiple countries and languages. It involves hreflang tags, country-specific domains or subfolders, localized content, and understanding how search behavior differs by region. Not just translation, but adaptation.
Tracking what products you have in stock, what's running low, and what needs reordering. For e-commerce, this also means syncing inventory across sales channels (your website, Amazon, physical store) so you don't oversell. Good inventory management prevents stockouts (losing sales) and overstock (tying up cash in unsold products). Most e-commerce platforms have built-in tools, or you can use dedicated software like TradeGecko or Cin7.
A UI framework for building cross-platform mobile apps using web technologies. It provides a library of mobile-optimized components that look native on both iOS and Android. Paired with Capacitor, it's a solid option for teams that know web dev but not mobile.
A peer-to-peer network for storing and sharing files in a distributed way. Instead of downloading from one server, you grab pieces from multiple nodes. Content is addressed by what it contains, not where it's located, so files can't be censored or lost if one server goes down.
The best of both worlds - pages are static but can rebuild themselves in the background when data changes. You get static speed with dynamic freshness.
The programming language of the web. Every browser runs it, and it's grown from a simple scripting tool to powering full applications on both the frontend and backend (via Node.js). Love it or hate it, you can't escape it in web development.
Android's modern UI toolkit that uses a declarative approach to building interfaces in Kotlin. It's Google's answer to SwiftUI. Instead of XML layouts, you write Kotlin functions that describe your UI. Simpler code, live previews, and less boilerplate.
A compact, self-contained token that carries user info and permissions. The server doesn't need to store session data because everything's in the token.
A visual workflow system using columns like "To Do," "In Progress," and "Done." Work flows continuously without fixed sprints. Great for teams with ongoing tasks.
When multiple pages on your site compete for the same keyword, confusing search engines about which one to rank. Instead of one strong page, you end up with two weak ones splitting the ranking power. Fix it by consolidating or differentiating the content.
A score estimating how hard it'll be to rank on page one for a specific keyword. Higher difficulty means more authoritative sites are competing. New sites should target low-difficulty keywords first and work their way up as their authority grows.
The words and phrases people type into search engines when they're looking for something. Your job is to figure out which ones your audience uses and make sure your content shows up for them.
Training a smaller "student" model to replicate the behavior of a larger "teacher" model. The student doesn't just learn from data, it learns from how the teacher responds to data. You end up with a compact model that punches above its weight.
The info box Google shows on the right side of search results for well-known entities like companies, people, and places. It pulls data from various sources including Wikipedia, your website, and Google Business Profile. Getting one boosts credibility.
Google's preferred language for Android development. It's more concise and safer than Java, with features like null safety and coroutines built in. Google declared it the official Android language in 2019, and new Android projects almost always use it.
A measurable value that shows whether you're achieving a business objective. For a website, KPIs might include monthly organic traffic, conversion rate, cost per lead, and revenue from web leads. The key word is "key." Having 50 KPIs defeats the purpose. Pick 3 to 5 metrics that directly reflect your business goals and track those religiously. Everything else is supporting detail.
An orchestration platform that manages containers at scale. It handles deployment, scaling, and recovery automatically. Often called K8s.
AWS's serverless computing service that runs your code in response to events without provisioning servers. You upload a function, set a trigger (like an API call or file upload), and AWS handles everything else. You only pay for the milliseconds your code actually runs.
A standalone page designed for a single purpose, usually to convert visitors from an ad or email campaign. It strips away navigation and distractions, focusing entirely on one call to action. Good landing pages can convert at 5-10%, while homepages average 1-2%.
A standalone web page created for a specific marketing campaign. Unlike your homepage (which tries to serve everyone), a landing page has one goal and removes all distractions. No navigation menu, no sidebar, no links to your blog. Just a headline, a value proposition, social proof, and a conversion form. Good landing pages convert 5 to 15% of visitors. Bad homepages convert 1 to 2%.
How long you wait between sending a request and getting the first response back. Lower latency means snappier interactions. Users notice anything over 200ms.
Loading content only when it's needed—like images that load as you scroll to them, or code that loads when you navigate to a page.
How long it takes for the biggest visible element on your page to load—usually a hero image or headline. Google wants this under 2.5 seconds.
A free resource you offer in exchange for someone's email address. Could be an ebook, checklist, template, or free tool. The idea is simple: give something valuable away to start a relationship with a potential customer. Good lead magnets solve a specific problem and attract the right audience. Bad ones attract freebie hunters who'll never buy.
The process of building relationships with potential customers who aren't ready to buy yet. Through email sequences, helpful content, and timely follow-ups, you stay top of mind until they're ready to make a decision. Most leads don't convert on the first visit. Lead nurturing bridges the gap between initial interest and eventual purchase. Companies that nurture leads generate 50% more sales-ready leads at 33% lower cost.
The legal terms that define how software can be used, distributed, or modified. Read the fine print - some "free" tools have strict commercial restrictions.
The practice of getting other websites to link to yours to improve search rankings. Quality matters more than quantity. One link from the New York Times is worth more than 1,000 links from random blogs. It's the hardest part of SEO.
The AI brain behind ChatGPT and similar tools. It's a massive program trained on tons of text that can understand and generate human-like writing. Think of it as autocomplete on steroids.
A system that distributes incoming traffic across multiple servers. Prevents any single server from getting overwhelmed and improves reliability.
Studying your server logs to see exactly which pages Googlebot crawls, how often, and what responses it gets. It's the most accurate way to understand how Google interacts with your site. Most SEOs skip this, which means most SEOs miss critical issues.
Recording events and data from your application as text entries. Good logs tell you what happened, when, and why. They're your forensic evidence when things go wrong. Structured logging (JSON format) makes them searchable and analyzable.
Longer, more specific search phrases that fewer people search for but are way easier to rank for. They usually convert better because the searcher knows exactly what they want.
A lightweight fine-tuning technique that trains only a tiny fraction of a model's parameters instead of the whole thing. You can customize a massive model on a single GPU that would normally need a server farm. It's become the go-to method for affordable model customization.
The total revenue you expect from a single customer over their entire relationship with you. If customers pay $100/month and stay 2 years on average, LTV is $2,400. The golden rule: LTV should be at least 3x your CAC, or you're losing money.
Teaching computers to learn patterns from data instead of programming explicit rules. You show it examples, it figures out the patterns. The umbrella term for most modern AI.
The ongoing support phase after launch. Bug fixes, security updates, small improvements. Your software doesn't end at launch - it needs care.
Using software to automate repetitive marketing tasks. Email sequences that send automatically based on triggers. Social media posts that schedule themselves. Lead scoring that prioritizes your hottest prospects without manual review. Tools like HubSpot, Mailchimp, and ActiveCampaign handle this. The goal is to nurture leads at scale without manually following up with each one.
The journey someone takes from discovering your business to becoming a customer. Usually described as stages: awareness (they learn you exist), consideration (they evaluate you against alternatives), and decision (they buy or don't). Each stage needs different content and messaging. Most businesses only focus on the bottom of the funnel and wonder why nobody's buying.
A tiny piece of code you put on your website that tracks visitor behavior and sends data back to ad platforms. The Meta Pixel tells Facebook who visited your site so you can retarget them. The Google tag does the same for Google Ads. Without pixels installed, you can't track conversions or build retargeting audiences. It's the first thing that should go on any business website.
A platform where multiple sellers list products for customers to buy. Amazon, Etsy, and eBay are the biggest examples. Selling on marketplaces gives you access to their existing traffic and customer base. The trade-offs are platform fees (typically 10-15% per sale), less control over branding and customer relationships, and competition from other sellers including the marketplace itself.
A standard way to connect AI models to external tools and data sources. It's like USB for AI—a universal plug that lets different systems talk to each other.
A system where services drop messages into a queue and other services pick them up asynchronously. It decouples your services so a slow email sender doesn't block your fast payment processor. RabbitMQ, SQS, and Kafka are popular options.
The short summary that appears under your title in search results. It doesn't directly affect rankings, but a good one can dramatically improve your click-through rate.
HTML snippets that give search engines info about your page. They don't show up on the page itself but help with SEO and control how your page appears in search results.
An architecture where your app is split into small, independent services that communicate over APIs. Easier to scale and update individual pieces.
A file that describes a change to your database structure—like adding a new column or creating a table. Migrations let you version control your database.
A major checkpoint in the project timeline. Usually tied to payments and deliverables. Hit the milestone, get paid, move to the next phase.
An architecture where a model has multiple specialized "expert" sub-networks and a router that picks which ones to activate for each input. Only a fraction of the model runs at any time, so you get a huge model's quality at a fraction of the compute cost.
Tracking how users interact with your app, including screen views, button taps, crashes, and session duration. Firebase Analytics and Mixpanel are the go-to tools. Without analytics, you're making decisions based on guesses instead of data.
Designing for phones first, then adding complexity for larger screens. Since most users are on mobile, you start with the smallest screen and scale up.
A site that works well on phones—readable text, tappable buttons, no horizontal scrolling. Google uses mobile-first indexing, so if your mobile experience sucks, your rankings will too.
A high-fidelity visual design showing exactly how the final product will look. Unlike wireframes, mockups include colors, typography, and real content.
Documentation that describes an AI model's capabilities, limitations, training data, and intended use cases. It's like a nutrition label for AI, helping developers understand what they're working with before they build on it.
Training a smaller, faster model to mimic a larger model's outputs. You get most of the performance at a fraction of the cost and speed.
Running an AI model on a server so applications can send it requests and get predictions back. It's the infrastructure that makes AI actually usable in production.
Tracking your app's health, performance, and errors in real-time. You can't fix what you can't see—monitoring tells you when things break before users complain.
A traditional architecture where the entire app is one big codebase and deployment unit. Simpler to start with, harder to scale later.
Keeping multiple related projects in a single Git repository instead of separate repos. Your frontend, backend, and shared libraries all live together, making cross-project changes atomic. Google, Meta, and Microsoft use monorepos for their massive codebases.
The traditional web model where each page is a separate HTML document. Click a link, server sends a new page. Simple, reliable, and SEO-friendly out of the box.
The predictable revenue a subscription business earns every month. If you have 100 customers paying $50/month, your MRR is $5,000. It's the heartbeat metric for SaaS companies. Investors look at MRR growth rate more than almost anything else.
The predictable revenue your business earns every month from active subscriptions. If you have 100 customers paying $50/month, your MRR is $5,000. MRR is the most important metric for subscription businesses because it's predictable and compounds. Tracking MRR growth, churn MRR (lost revenue from cancellations), and expansion MRR (revenue from upgrades) tells you whether your business is healthy.
The umbrella contract that covers the general terms of our working relationship. Individual projects get their own SOWs under this agreement.
Selling products across multiple platforms simultaneously: your own website, Amazon, eBay, Instagram Shop, retail stores. Each channel reaches different customers. The challenge is keeping inventory synced, maintaining consistent pricing, and managing orders from multiple sources. Multi-channel selling can significantly increase revenue but requires systems to prevent overselling and operational chaos.
Using multiple cloud providers (like AWS plus Google Cloud) to avoid vendor lock-in or meet specific requirements. The theory is great but the practice is complicated. Most companies that claim multi-cloud really just use one primary provider with a backup.
AI that can work with multiple types of input like text, images, audio, and video all at once. Instead of being text-only or image-only, it understands them together.
An AI model that can process and generate multiple types of content, like text, images, audio, and video, within the same system. GPT-4o and Gemini are multimodal because you can show them a photo and ask questions about it.
The simplest version of your product that still solves the core problem. Ship fast, learn from real users, then improve. Don't build everything at once.
AI that identifies and classifies proper nouns and specific items in text, like people, companies, dates, and locations. It can scan a news article and pull out every person mentioned, every company named, and every date referenced.
An app built specifically for one platform (iOS or Android) using that platform's native language. It's faster and can use all device features, but you have to build it twice.
AI that understands and works with human language. It's how chatbots understand your questions, how spam filters catch junk mail, and how voice assistants know what you're asking for.
A contract that says we won't share your confidential information. Standard practice before discussing proprietary ideas or business secrets.
Revenue from existing customers this year compared to what they paid last year, including upgrades and churn. Over 100% means existing customers are spending more over time even accounting for cancellations. Best SaaS companies hit 120-140% NRR.
Software loosely inspired by how brains work—layers of connected "neurons" that process information. It's the foundation of modern AI, though it's really just math, not actual brain stuff.
A React framework by Vercel that adds server-side rendering, routing, and API endpoints out of the box. It's become the default way to build production React apps because it handles all the hard stuff, like SEO, performance, and deployment.
A runtime that lets you run JavaScript outside the browser, primarily on servers. It enabled full-stack JavaScript development, meaning you can write both your frontend and backend in the same language. Powers Express, Next.js API routes, and millions of servers.
A link with a special tag telling search engines not to pass ranking authority to the linked page. Used for paid links, user-generated content, and untrusted sources. Google introduced it to combat link spam in blog comments and forums.
Databases that don't use traditional tables and SQL. They're more flexible—great for unstructured data or when you need to scale massively.
A standard that lets you log into apps using your Google, GitHub, or other accounts without sharing your password with each app.
AI that identifies and locates specific things in images or video, drawing bounding boxes around them. It doesn't just say "there's a car," it shows you exactly where the car is. Used in security cameras, self-driving cars, and retail analytics.
The ability to understand what's happening inside your system by looking at its outputs: logs, metrics, and traces. Monitoring tells you something is broken. Observability helps you figure out why. It's the difference between an alarm and a diagnosis.
Designing mobile apps to work without an internet connection by default, syncing when connectivity returns. It stores data locally and handles conflicts when the server and device have different versions. Essential for apps used in areas with spotty coverage.
The percentage of email recipients who open your email. If you send to 1,000 people and 250 open it, your open rate is 25%. Industry averages range from 15 to 25%. Subject lines drive open rates more than anything else. A compelling subject line can double your open rate overnight. Note: Apple's Mail Privacy Protection inflates open rates for Apple users, so take these numbers with a grain of salt.
Software with publicly available source code that anyone can use, modify, and distribute. Free to use doesn't always mean free of obligations though.
AI models where the code and often the weights are publicly available. You can download them, modify them, and run them yourself without paying per API call.
The percentage of people who click your result after seeing it in search. Position 1 gets about 30% of clicks, position 5 gets around 5%. Improving your title tag and meta description can boost CTR without improving rankings, getting you more traffic for free.
Visitors who find your site through unpaid search results. They clicked on your link because Google thought your content matched what they were looking for—not because you paid for the spot.
Visitors who find your website through unpaid search results. They Googled something, your page appeared, they clicked. No ad spend involved. Organic traffic is the result of SEO work: optimizing your site, creating content, building authority. It's the most valuable traffic type because it's free and ongoing. Unlike paid traffic, organic traffic keeps coming even after you stop actively working on it.
A tool that lets you work with your database using regular code instead of writing SQL. You work with objects, and it translates to database queries.
Pushing updates to your mobile app without going through the App Store or Google Play review process. React Native and Expo apps can update JavaScript bundles instantly. It's how you push a critical bug fix in minutes instead of waiting days for store approval.
Similar to domain authority, but for a single page instead of your whole site. Some pages on your site might be stronger than others based on their backlinks and content quality.
Google's umbrella ranking signal combining Core Web Vitals, mobile-friendliness, HTTPS, and intrusive interstitial guidelines. It measures whether your pages are pleasant to use, not just whether they have good content. A great article on a terrible website won't rank well.
How fast your website loads. Faster sites rank better and convert more. Every second of delay costs you visitors and money.
A 0-100 rating of how fast your web page loads, measured by tools like Google PageSpeed Insights. Scores above 90 are considered good, 50-89 needs improvement, and below 50 is poor. The score factors in loading time, interactivity, and visual stability. For business websites, aim for 90+ on mobile. Desktop scores are usually higher and less meaningful since mobile is where most traffic happens.
A count of every time a page on your website is loaded. If one visitor views 5 pages, that's 5 pageviews. Pageviews alone don't tell you much. 10,000 pageviews sounds impressive until you realize 9,000 of them bounced. Always pair pageviews with engagement metrics like time on page and conversion rate to get the full picture of how your content performs.
Visitors who come to your site because you paid for an ad. This includes Google Ads, social media ads, or any other promotion where you're spending money to get clicks.
Visitors who come to your website through paid advertising. Google Ads, Meta Ads, LinkedIn Ads, display networks. You pay for each click or impression. The advantage is speed: you can drive traffic today. The disadvantage is cost: it stops the moment you stop paying. Smart businesses use paid traffic to generate revenue while building organic traffic for the long term.
Publishing content on high-authority domains like Reddit, Medium, or LinkedIn to rank quickly instead of building your own site's authority. It works because these platforms already have massive domain authority. Google has been cracking down on abusive uses.
Google's ability to rank a specific passage within a longer page for a query, even if the overall page isn't focused on that topic. A single paragraph in a 3,000-word article can rank on its own if it perfectly answers a question.
A password replacement that uses your device's biometrics (fingerprint or face) to log you in. No passwords to remember, nothing to type, and phishing-resistant by design. Apple, Google, and Microsoft are all pushing passkeys as the future of login.
The service that processes credit card and digital payments on your website. Stripe, PayPal, Square are common examples. The gateway securely transmits payment information between your customer, your site, and the bank. Choosing the right one affects checkout experience, transaction fees (typically 2.9% + $0.30 per transaction), supported payment methods, and international capabilities.
A network where computers communicate directly with each other instead of going through a central server. BitTorrent, blockchain networks, and WebRTC calls are all peer-to-peer. Each participant is both a client and a server.
Hiring security experts to try to hack your systems, then telling you what they found. It's a controlled attack to discover vulnerabilities before real hackers do. Also called "pen testing." Better to pay a friendly hacker than get surprised by a hostile one.
A Google SERP feature showing related questions that expand to reveal short answers. Getting your content featured here drives extra traffic. The trick is structuring your content to directly answer common questions with clear, concise paragraphs.
A measurement of how surprised a language model is by new text. Lower perplexity means the model predicts text well. It's one of the standard ways to compare language models, though it doesn't always correlate with real-world usefulness.
A long, authoritative page that covers a broad topic at a high level and links out to detailed cluster pages on each subtopic. It's the hub of a content cluster. Think of it as the table of contents for everything your site says about a subject.
A strategy where the product itself drives acquisition, retention, and expansion instead of sales teams. Users sign up, experience the product's value, and upgrade on their own. Slack, Figma, and Notion are classic PLG success stories.
The "Sean Ellis test" for product-market fit. You ask users "How would you feel if you could no longer use this product?" If 40% or more say "very disappointed," you've likely found product-market fit. Below 40% means you need to keep iterating.
A small test to prove an idea actually works before investing serious time or money. It's not a product—it's evidence that building the product makes sense.
An advertising model where you pay only when someone clicks your ad. Google Ads is the most common PPC platform. You bid on keywords, set a daily budget, and pay each time a searcher clicks through to your site. PPC gives you immediate visibility and precise control over targeting. But it requires constant optimization. Set it and forget it is the fastest way to waste an ad budget.
Offering multiple versions of your product at different price points, usually called Free, Pro, and Enterprise (or similar). Each tier adds more features. Three tiers work best because the middle option becomes the most popular choice, a psychological pricing trick.
A TypeScript-first ORM that makes database queries type-safe and auto-completed. You define your schema in a readable format, and Prisma generates a client with full TypeScript types. It makes database mistakes harder to make and easier to catch.
A structured file containing all your product data: names, prices, descriptions, images, availability. Product feeds are used by Google Shopping, Facebook Marketplace, and comparison shopping engines to list your products. Keeping your feed accurate and up-to-date is critical. Incorrect pricing or out-of-stock items in your feed waste ad money and frustrate customers.
Improving your product pages to increase sales. Good product pages have clear photos from multiple angles, detailed descriptions that address buyer concerns, visible pricing, social proof (reviews and ratings), clear add-to-cart buttons, and shipping information. The product page is where buying decisions happen. Every element should either build confidence or reduce friction.
Images of your products that help customers understand what they're buying. Online shoppers can't touch or try products, so photos do all the heavy lifting. Include multiple angles, lifestyle shots (product in use), close-up details, and size references. White background photos are standard for product listings. Lifestyle photos work better for social media and ads. Poor product photos are the fastest way to kill conversions.
When your product solves a real problem well enough that people actively want it and tell others about it. You know you have it when growth feels natural instead of forced. Without product-market fit, everything else, marketing, hiring, fundraising, is premature.
Creating thousands of pages at scale using templates and data, targeting long-tail keywords automatically. Zapier has a page for every app integration, Yelp has one for every business. It works when you have structured data that matches real search demand.
The art of writing instructions that get AI to do what you actually want. It's surprisingly important—the same AI can give garbage or gold depending on how you ask.
A security attack where someone tricks an AI into ignoring its instructions by hiding malicious prompts in the input. It's the AI equivalent of SQL injection.
AI models owned by a company that you can only access through their API or products. You don't get the model weights or full control over how it runs.
Data you pass into a component to customize it. Short for "properties." They're how parent components talk to their children.
A working model of your product used for testing and feedback. It doesn't need to be pretty or complete—it needs to let people interact with your idea.
An Apple-specific file that ties your app, your developer account, and specific devices together. It's basically permission to run your app.
Messages sent directly to users' phones even when they're not in your app. Powerful for engagement, but abuse them and users will uninstall.
A website that acts like a native app - works offline, sends push notifications, and can be installed on your home screen. All the app benefits without the app store.
Google's rating of the quality and relevance of your ads and landing pages, scored 1 to 10. Higher quality scores mean you pay less per click and get better ad positions. It's based on three factors: expected click-through rate, ad relevance, and landing page experience. Improving your quality score is one of the fastest ways to reduce ad costs without cutting budget.
Shrinking an AI model by using less precise numbers for its weights. A 4-bit quantized model is way smaller and faster than the original, with only a small quality drop.
A technique that lets AI search your documents before answering questions. Instead of just making stuff up, it pulls real info from your data first. This is how you build a chatbot that actually knows your business.
Capping how many requests someone can make in a time period. It protects your server from abuse and ensures fair usage.
Restricting how many requests a user or IP can make in a given time window to prevent abuse. It stops brute force password attacks, API scraping, and DDoS attempts. If someone tries 100 passwords in a minute, rate limiting locks them out.
Meta's JavaScript library for building user interfaces using reusable components. It's the most popular frontend framework by a wide margin, powering Facebook, Instagram, Netflix, and thousands of other apps. The virtual DOM and component model changed how we build web UIs.
Facebook's framework for building cross-platform mobile apps using JavaScript and React. You write once and it compiles to native iOS and Android code.
Seeing what's happening on your website right now. How many people are on the site, which pages they're viewing, where they came from. Useful during product launches, ad campaigns, or events to see immediate impact. Less useful for regular analysis because real-time data is noisy and encourages reactive rather than strategic decisions. Check it during campaigns, ignore it otherwise.
AI that suggests things you might like based on your behavior and preferences. Netflix recommending shows, Spotify suggesting songs, and Amazon showing "customers also bought" all run on recommendation systems. They're responsible for a huge chunk of online revenue.
Teaching AI through trial and error with rewards and penalties. The model tries different actions, gets feedback on what worked, and gradually learns the best strategy. It's how AlphaGo learned to beat world champions at Go and how robots learn to walk.
A React framework that embraces web standards like forms, HTTP caching, and progressive enhancement. It focuses on server-side data loading and mutations, making apps work even before JavaScript loads. Now maintained by Shopify.
The resources Google allocates to rendering JavaScript-heavy pages during crawling. If your site relies heavily on client-side JavaScript, Google might not render all your content, meaning it won't index what it can't see. Server-side rendering fixes this.
The process of documenting exactly what the software needs to do. We ask a lot of questions now so we don't build the wrong thing later.
Building websites that look good on any screen size—phone, tablet, or desktop. The layout adjusts automatically based on the browser width.
Serving different image sizes based on the user's screen resolution and size. A phone doesn't need a 4K image designed for a desktop monitor. Using srcset and the picture element saves bandwidth and makes pages load significantly faster on mobile.
The most common type of API that uses standard web requests like GET, POST, PUT, and DELETE. It's simple, predictable, and works everywhere.
A monthly agreement where you pay a set amount for ongoing work or availability. Good for continuous development needs. You get priority, they get predictable income.
Showing ads to people who've already visited your website or interacted with your business. Someone looks at your pricing page, leaves without buying, then sees your ad on Instagram the next day. That's retargeting. It works because these people already know you exist and showed interest. Retargeting typically has the highest ROI of any ad type because the audience is already warm.
The percentage of customers who continue doing business with you over a given period. The flip side of churn rate. If 95 out of 100 customers stick around, your retention rate is 95%. High retention is the foundation of sustainable growth. A 5% increase in retention can increase profits by 25 to 95% according to Harvard Business School research. Focus on keeping customers happy, not just acquiring new ones.
The full name for RAG. It combines information retrieval (searching a knowledge base) with text generation (the LLM writing a response). The retrieval step grounds the AI in real facts, which dramatically reduces hallucinations.
Your rules for how customers can return products and get refunds. A generous return policy actually increases sales because it reduces purchase anxiety. Zappos built their business partly on free returns. Display your return policy prominently on product pages and during checkout. Common terms to define: return window (14-30 days typical), who pays return shipping, refund vs store credit, and condition requirements.
An AI model trained to score how good another model's outputs are, based on human preferences. It acts as an automated judge during RLHF training, telling the main model which responses humans would prefer.
A document companies send out when they want agencies to bid on a project. It describes what they need and asks for your approach and pricing.
Enhanced search results that show extra info like star ratings, prices, or recipe times. They make your listing stand out and get more clicks.
Reinforcement Learning from Human Feedback. Humans rate AI outputs as good or bad, and that feedback trains the model to give better answers. It's how ChatGPT learned to be helpful and safe.
How much revenue you earn for every dollar spent on advertising. A 4x ROAS means you make $4 for every $1 in ad spend. Different industries have different benchmarks, but most businesses should target at least 3x to be profitable after accounting for product costs and overhead. ROAS below 2x usually means the campaign needs work or should be paused.
A text file that tells search engine bots which pages they can and can't crawl on your site. It's like a bouncer at a club—you decide who gets in where.
How much profit you made relative to what you spent. If you invest $1,000 in marketing and generate $5,000 in revenue (with $2,000 in costs), your ROI is 200%. The formula is (Revenue - Cost) / Cost x 100. Every marketing dollar should be evaluated against ROI. If something isn't generating a positive return, it needs to be fixed or stopped. Simple in theory, surprisingly hard to measure accurately.
Reverting to a previous working version when something goes wrong. It's your "undo" button for deployments—essential when updates break things.
How many months a startup can survive before running out of money at the current burn rate. Cash in bank divided by monthly burn. If you have $500K and burn $50K/month, you have 10 months of runway. Below 6 months and panic mode kicks in.
AWS's object storage service that's become the de facto standard for storing files in the cloud. Images, backups, logs, static websites—it all goes in S3. It's cheap, virtually unlimited, and 99.999999999% durable (that's eleven nines).
Software you access through the internet and pay for on a subscription basis. You don't install anything - just log in and use it.
The structure of your database—what tables exist, what columns they have, and how they relate to each other. It's like a blueprint.
A specific vocabulary of structured data tags that help search engines understand your content. You add JSON-LD code to your pages telling Google exactly what type of content it is, like a product, recipe, event, or FAQ. It's the key to earning rich snippets.
When a project keeps growing beyond the original plan because people keep adding "just one more thing." The silent killer of timelines and budgets.
How far down a page visitors scroll before leaving. If most people only see the top 25% of your page, everything below that mark is invisible to them. This matters because important content, CTAs, or social proof buried below the scroll point might as well not exist. Track scroll depth to ensure your most important elements are placed where people actually see them.
A specific agile framework with defined roles, sprints, and ceremonies like daily standups and retrospectives. More structured than just "being agile."
The reason behind a search query. Someone searching "best coffee maker" wants to buy (commercial intent), while "how does a coffee maker work" wants to learn (informational intent). Matching your content to the right intent is critical for ranking.
Securely storing and accessing passwords, API keys, certificates, and other sensitive data. Hardcoding secrets in your codebase is a security nightmare. Tools like AWS Secrets Manager, HashiCorp Vault, and Doppler encrypt and rotate them properly.
A thorough review of your systems, code, and practices to find security weaknesses. It covers everything from code vulnerabilities to employee access policies to how you handle backups. Think of it as a health checkup for your security posture.
A specific type of attention where each word in a sentence looks at every other word to understand relationships. It's how the model figures out that "it" in "The cat sat on the mat because it was tired" refers to the cat, not the mat.
Search that understands meaning, not just keywords. It finds results based on what you meant, even if you used different words. Way smarter than traditional search.
AI that determines whether text expresses positive, negative, or neutral feelings. Businesses use it to monitor brand perception, analyze customer reviews, and gauge social media reactions at scale without reading every single comment.
The page you see after typing something into Google. It's where all the search results live—ads at the top, organic results below, and sometimes featured snippets, maps, or images mixed in.
Special result types on Google beyond standard blue links. These include featured snippets, knowledge panels, image packs, video carousels, People Also Ask boxes, and local map packs. They take up more screen space and can steal clicks from organic results.
React components that run only on the server and send just the HTML to the browser, with zero JavaScript shipped to the client. They can directly access databases and APIs. Introduced by React and popularized by Next.js App Router.
Backend code that runs on-demand without you managing servers. You write the function, cloud provider handles scaling. Pay only when it executes.
A dedicated infrastructure layer that handles communication between microservices. It manages traffic routing, load balancing, encryption, and observability without changing your application code. Istio and Linkerd are the main options.
A background script that sits between your web app and the network. It can intercept requests, cache responses, and enable offline functionality. Service workers are what make progressive web apps work without an internet connection.
A single visit to your website. A session starts when someone arrives and ends after 30 minutes of inactivity or at midnight. One person can generate multiple sessions. If they visit your site in the morning, leave, and come back in the afternoon, that's two sessions from one user. Sessions are one of the most basic analytics metrics and the foundation for calculating rates like bounce rate and conversion rate.
Keeping track of logged-in users between requests. Since HTTP is stateless, you need sessions or tokens to remember who's who.
Recordings of real user sessions showing exactly how they navigate your site, including mouse movements, clicks, scrolling, and rage clicks. Like watching over someone's shoulder while they use your site. Tools like Hotjar and FullStory make this easy.
A replay of an actual visitor's experience on your website. You can watch them scroll, click, hesitate, and navigate. It's like looking over their shoulder. Session recordings expose usability issues that data alone can't catch: confusing navigation, broken forms, elements that look clickable but aren't. Microsoft Clarity offers this for free. Five minutes of watching recordings will show you problems you never knew existed.
A way to encapsulate a component's HTML and CSS so it doesn't leak into or get affected by the rest of the page. Your component's styles stay contained, solving the age-old problem of CSS conflicts in large projects.
Your approach to getting products to customers and what you charge for it. Free shipping increases conversion rates but eats into margins. Flat-rate shipping is predictable for customers. Real-time calculated shipping is most accurate but can shock customers with high costs. Many successful stores build shipping costs into product prices and offer "free" shipping. The right strategy depends on your margins, product weight, and competition.
The most popular hosted e-commerce platform. Shopify handles hosting, security, payments, and provides a store builder with themes and apps. Plans start at $39/month plus transaction fees. It's excellent for small to mid-size stores that want to start selling quickly without managing infrastructure. Limitations show up as you scale: customization constraints, app dependency, and transaction fees that add up at high volume.
When someone adds items to their online cart but leaves without completing the purchase. The average abandonment rate is around 70%, meaning 7 out of 10 people who start buying don't finish. Common causes: unexpected shipping costs, complicated checkout, required account creation, and slow loading. Reducing abandonment by even a few percent can significantly increase revenue without driving any additional traffic.
Formal approval that something is done and accepted. Once you sign off, we move forward. It protects both of us from endless revision loops.
The practice of making your website load faster through techniques like image compression, code minification, lazy loading, and caching. Every 100ms of improvement impacts conversion rates. Google PageSpeed Insights and Lighthouse are the standard testing tools.
An XML file that lists all the important pages on your site. It helps search engines discover and crawl your content faster, especially useful for new or large sites.
A unique code assigned to each product variant in your inventory. A blue t-shirt in size medium has a different SKU than the same shirt in large. SKUs help you track inventory, analyze which variants sell best, and manage reordering. They're different from barcodes (which are standardized externally). SKUs are your internal system for organizing products. The more products you have, the more critical a clear SKU system becomes.
Service Level Agreement—a contract promising specific performance levels like uptime or response time. If they miss it, you usually get credits.
Programs stored on a blockchain that automatically execute when conditions are met. They're called "contracts" because they enforce agreements without middlemen. Once deployed, nobody can change them, which is both a strength and a risk.
Evidence that other people trust your business. Reviews, testimonials, client logos, case studies, social media follower counts, "as seen in" badges. Humans are wired to follow the crowd. If 500 people gave your business 5 stars, a new visitor assumes you must be good. Social proof is the most effective trust-building element on any website or landing page.
A document that spells out exactly what we're building, how long it'll take, and how much it'll cost. The project blueprint that keeps everyone aligned.
A web app that loads once and never does a full page refresh. Navigation feels instant because it just swaps content instead of requesting new HTML from the server.
A speed trick where a small, fast model drafts several tokens ahead, then the large model checks them all at once. It's like having an intern write a rough draft that the expert just approves or corrects. Can make inference 2-3x faster.
AI that converts spoken words into text. It's the technology behind Siri, Alexa, and meeting transcription tools. Modern systems can handle different accents, background noise, and even multiple speakers talking at once.
A fixed time period (usually 1-2 weeks) where your team commits to completing specific work. At the end, you ship something usable and plan the next sprint.
A language for talking to databases. You use it to ask questions like "give me all users who signed up this week" or "update this user's email address."
An attack where malicious SQL code gets inserted into input fields to manipulate your database. An attacker could type something into a login form that tricks your database into handing over all user data. Parameterized queries prevent this.
A security certificate that encrypts data between your website and your visitors. It's what puts the "S" in HTTPS and shows the lock icon in the browser. For e-commerce, SSL is non-negotiable. Without it, browsers display "Not Secure" warnings, payment processors won't work, and customers won't trust you with their credit card. Most hosting providers include SSL for free through Let's Encrypt.
Encryption protocols that secure data between browser and server. TLS is the modern version, but everyone still calls it SSL.
The server builds the full HTML page before sending it to the browser. Great for SEO and fast initial loads because users see content immediately instead of a blank screen.
Anyone who has a say in or is affected by the project. This includes decision-makers, end users, and anyone who'll complain if it goes wrong.
How you handle data that changes over time in your app—like whether a user is logged in, what's in their cart, or if a modal is open.
Building all your pages at deploy time so they're just HTML files sitting on a CDN. Blazing fast and dirt cheap to host since there's no server computing anything.
Getting AI responses word-by-word as they're generated instead of waiting for the complete answer. It makes apps feel faster because users see text appearing in real-time.
Server-side rendering that sends HTML in chunks as it's ready instead of waiting for the entire page to render. Users see content loading progressively instead of staring at a blank screen. It makes slow pages feel much faster.
Code you add to your pages that helps search engines understand your content better. It can get you rich results like star ratings, recipe cards, or FAQ dropdowns in search.
Getting AI to return data in a specific format like JSON instead of free-form text. Essential for building real applications because your code needs predictable, parseable responses, not prose it has to interpret.
A prefix added to your main domain that acts as a separate address. Useful for organizing different parts of your site.
A business model where customers pay a recurring fee for products or services. Monthly boxes, software subscriptions, membership sites, replenishment models (auto-shipping consumables). The appeal is predictable recurring revenue and higher customer lifetime value. The challenge is churn. Subscription businesses live and die by their ability to keep customers subscribed month after month.
Compromising software by attacking its dependencies or build process instead of the software itself. If a hacker inserts malicious code into a popular npm package, everyone who installs it gets infected. The SolarWinds hack is the most famous example.
A frontend framework that compiles your code into tiny, efficient vanilla JavaScript at build time. There's no virtual DOM or runtime framework code shipped to the browser. The result is smaller bundle sizes and faster apps with less boilerplate code.
The official application framework for Svelte, similar to what Next.js is for React. It handles routing, server-side rendering, and deployment. If you're building a production Svelte app, you're using SvelteKit.
Apple's programming language for building iOS, macOS, and watchOS apps. It replaced Objective-C with a cleaner, safer, and more modern syntax. If you're building a native iOS app, you're writing Swift.
Apple's modern framework for building interfaces across iPhone, iPad, Mac, and Apple Watch using the Swift language. It uses a declarative syntax where you describe what you want and SwiftUI figures out how to render it. The future of Apple-platform development.
Fake data generated by AI to train other AI models. When you don't have enough real data or it's too sensitive to use, synthetic data can fill the gap.
Hidden instructions given to an AI model before the user conversation starts. It sets the model's behavior, personality, and constraints. When ChatGPT acts like a helpful assistant instead of a base language model, that's the system prompt at work.
A utility-first CSS framework where you style elements using pre-built class names like "bg-blue-500" and "text-center" directly in your HTML. It feels weird at first but dramatically speeds up development once it clicks. You rarely write custom CSS.
Shortcuts in code that save time now but create more work later. Like credit card debt—sometimes worth it for speed, but it accumulates interest if you don't pay it down.
A setting that controls how creative or random AI responses are. Low temperature = predictable and focused. High temperature = creative but potentially weird.
The most popular infrastructure as code tool, made by HashiCorp. You write configuration files describing your servers, databases, and networking, then Terraform creates it all automatically. It works with AWS, Google Cloud, Azure, and dozens of other providers.
Apple's official beta testing platform for iOS apps. You upload a build, invite testers, and they can install it before it hits the App Store.
AI that sorts text into predefined categories. Spam detection, topic categorization, support ticket routing, and content moderation are all text classification. The model reads the text and decides which bucket it belongs in.
AI that converts written text into natural-sounding spoken audio. Modern TTS can clone specific voices, add emotion, and sound nearly indistinguishable from real humans. It powers audiobooks, virtual assistants, and accessibility tools.
How many requests your AI system can handle per second. High throughput means you can serve more users simultaneously without things grinding to a halt.
A billing model where you pay for actual hours worked plus any materials or software costs. Good for projects where scope isn't 100% clear upfront.
The clickable headline that shows up in search results. It's one of the most important on-page SEO factors—make it compelling and include your target keyword.
The maximum number of tokens an AI model can process in a single conversation, including both the input and output combined. If you hit the limit, the model starts "forgetting" earlier parts of the conversation. Bigger windows cost more to run.
The component that chops text into tokens before feeding it to an AI model. Different models use different tokenizers, which means the same sentence can be split differently. It's why token counts vary between GPT and Claude.
The chunks AI breaks text into for processing. Roughly 4 characters or 3/4 of a word. This is how AI pricing works—you pay per token in and out.
An AI's ability to use external tools like web search, calculators, code execution, or APIs to accomplish tasks. It extends what the AI can do beyond just generating text.
When search engines recognize your site as a trusted expert on a specific subject because you've published thorough, interlinked content covering all aspects of that topic. A site with 50 well-written articles about coffee will outrank one with 5.
Where your website visitors come from. The main categories are organic search (Google), paid search (Google Ads), social media (Facebook, Instagram, LinkedIn), direct (typing your URL or bookmarks), referral (links from other sites), and email. Knowing your traffic sources tells you which marketing channels work and where to invest more. A healthy site has traffic from multiple sources rather than depending on one.
The examples an AI learns from. Garbage in, garbage out—the quality of training data directly affects how good the AI is. It's the textbooks the AI studied.
Reusing a model trained on one task as the starting point for a different task. Instead of training from scratch, you take a model that already understands language or images and teach it your specific thing. It's why fine-tuning works so well.
The architecture that makes modern AI work. It's a specific way of building neural networks that's really good at understanding context and relationships in data. The "T" in GPT.
A TypeScript library that lets your frontend call backend functions with full type safety and no API layer in between. Change a function on the server, and your IDE immediately tells you what broke on the client. Eliminates entire categories of API bugs.
Vercel's Rust-based JavaScript bundler designed to replace Webpack, claiming to be up to 700x faster. It's built into Next.js and optimized for the React ecosystem. Still maturing, but it represents the trend of rewriting JS tooling in faster languages.
A build system for JavaScript monorepos that caches task results and runs things in parallel. If nothing changed in a package, it skips rebuilding it. Saves hours on CI builds for large projects. Made by Vercel.
Requiring two different types of proof to log in, usually a password plus a code from your phone. Even if someone steals your password, they can't get in without the second factor. Use an authenticator app, not SMS, because SIM swapping is real.
JavaScript with type checking. You declare what kind of data variables hold, and the compiler catches bugs before your code runs. It adds some upfront work but prevents entire categories of runtime errors. Most serious JavaScript projects use it now.
The final testing phase where actual users (or you) verify the software works as expected. It's your chance to catch issues before we call it done.
The revenue and costs associated with a single customer or unit. If you spend $50 to acquire a customer who generates $200 in profit over their lifetime, your unit economics are healthy. If those numbers are inverted, you're burning money with every sale.
Apple's system for opening specific app screens from regular web URLs. Tap a link to example.com/product/123, and if the user has the app, it opens there instead of the browser. Unlike custom URL schemes, they're regular HTTPS links that work either way.
Encouraging a customer to buy a more expensive version of what they're already considering. "Want the premium plan instead of basic?" or "Upgrade to the larger size for just $5 more." Done well, upselling increases average order value while genuinely giving the customer something better. Done poorly, it feels pushy and erodes trust. The best upsells offer clear additional value, not just a higher price tag.
The percentage of time your service is available and working. 99.9% uptime sounds great until you realize that's still 8+ hours of downtime per year.
Dividing your audience into groups based on shared characteristics. New visitors vs returning visitors. Mobile vs desktop users. Visitors from Google Ads vs organic search. Analyzing segments separately reveals insights hidden in aggregate data. Your overall conversion rate might be 3%, but new mobile users from Facebook might convert at 0.5% while returning desktop users convert at 15%. That changes your strategy completely.
Behavioral data that indicates how satisfied users are with search results, things like click-through rate, dwell time, and pogo-sticking (quickly bouncing between results). Google uses these signals to refine rankings over time.
A simple sentence describing what a user wants to do and why. Format: "As a [user], I want to [action] so that [benefit]." Keeps development focused on real needs.
Content created by your customers rather than your brand. Reviews, photos, videos, social media posts about your product. UGC is incredibly powerful because people trust other customers more than they trust your marketing. Encouraging and resharing UGC costs almost nothing and often outperforms professionally produced content in engagement and conversions.
Tags you add to URLs to track where your traffic comes from. They tell Google Analytics which campaign, source, and medium sent each visitor. Without UTMs, you're guessing which marketing efforts actually work.
Tags you add to URLs to track where traffic comes from. They look like ?utm_source=facebook&utm_medium=social&utm_campaign=spring-sale appended to your link. When someone clicks this tagged link, Google Analytics records exactly which campaign, source, and medium drove that visit. Without UTM tags, you're guessing which marketing efforts actually work. With them, you know exactly.
A special database built to store and search embeddings fast. Regular databases search by exact matches; vector databases find similar meanings. Essential for RAG and semantic search.
When switching away from a service becomes so painful or expensive that you're basically stuck with them. Something to avoid when possible.
Making an existing server more powerful by adding CPU, RAM, or storage. Simpler than horizontal scaling but has a ceiling. Eventually you hit the biggest server available. Databases often rely on vertical scaling because distributing them is hard.
A lightweight copy of the real DOM that React uses to figure out what actually changed. It only updates the parts that need updating, making things faster.
A modern build tool that starts your development server in milliseconds instead of seconds. It uses native ES modules during development so it doesn't need to bundle everything upfront. Created by the Vue.js creator but works with React, Svelte, and anything else.
A progressive JavaScript framework known for being easy to learn. It doesn't force you to adopt everything at once. You can use as much or as little as you need. Popular in Asia and with developers who find React's ecosystem overwhelming.
Automated tools that scan your systems for known security weaknesses. Unlike penetration testing, it's automated and doesn't try to actually exploit the vulnerabilities. It catches the low-hanging fruit like outdated software and misconfigurations.
A firewall specifically designed to protect web applications. It sits between users and your app, filtering out malicious requests like SQL injection attempts and XSS attacks. Cloudflare and AWS WAF are common choices.
A set of browser-native standards for creating reusable custom HTML elements. Unlike React or Vue components, web components work everywhere without a framework. You can create a <my-button> element that works in any project.
Background threads that let you run heavy JavaScript without freezing the user interface. Normally JavaScript is single-threaded, meaning complex calculations block everything else. Web workers handle the hard stuff in the background.
A vision for a decentralized internet where users own their data and identity through blockchain technology instead of relying on big tech companies. It includes cryptocurrencies, NFTs, and decentralized apps. Controversial because the tech is real but the hype often outpaces utility.
A binary format that lets you run code written in languages like C++, Rust, or Go directly in the browser at near-native speed. It's how Figma runs a full design tool in your browser without feeling slow, and how video editors work without installing software.
A JavaScript API that renders 2D and 3D graphics in the browser using your GPU. It's how browser games, 3D product viewers, and interactive data visualizations run without any plugins. Three.js is the most popular library built on top of it.
The next-generation graphics API replacing WebGL that gives browsers direct access to modern GPU features. It's significantly faster and can even run machine learning models in the browser. Think of it as WebGL but built for 2024 GPUs instead of 2011 ones.
A way for one app to notify another when something happens. Instead of constantly asking if something changed, you get pinged automatically.
The OG JavaScript bundler that combines all your files, images, and dependencies into optimized bundles for production. It's incredibly powerful but also incredibly complex to configure. Newer tools like Vite and Turbopack are gradually replacing it.
A technology that enables real-time audio, video, and data sharing directly between browsers without going through a server. It's what powers Google Meet, Discord voice chat, and peer-to-peer file sharing in the browser.
A persistent connection between browser and server that lets both sides send messages instantly. Unlike regular requests, the connection stays open.
Software built so you can rebrand it as your own. We build it, you slap your logo on it and sell it to your customers.
A basic sketch showing the layout and structure of a page without any design polish. Think blueprint, not finished building. It's about where things go, not how they look.
A free, open-source e-commerce plugin for WordPress. It turns any WordPress site into an online store. The plugin itself is free, but you'll pay for hosting, themes, and extensions. More customizable than Shopify since you control the code and server. But that flexibility comes with more responsibility: updates, security, performance, and hosting are all on you. Best for businesses already on WordPress or needing deep customization.
Apple's official IDE for developing iOS and macOS apps. It includes a code editor, simulator, debugger, Interface Builder, and everything else you need. Love it or hate it, there's no alternative for building and submitting apps to the App Store.
An attack where malicious JavaScript gets injected into a website and runs in other users' browsers. An attacker might sneak a script into a comment field that steals cookies or redirects users. It's one of the most common web vulnerabilities.
A security model that assumes nothing and nobody should be trusted by default, even inside your own network. Every request gets verified regardless of where it comes from. The old approach of "trust everything inside the firewall" doesn't work anymore.
A cryptographic method that lets you prove you know something without revealing the actual information. Like proving you're over 21 without showing your birthday, or proving you have enough money without revealing your balance. Used in privacy-focused systems.
When an AI can do a task it wasn't explicitly trained on, without any examples. You just describe what you want and it figures it out from its general knowledge.
491 terms across 10 categories