Web Development Process: 7 Stages Explained
The web development process is the sequence of stages a website goes through from first idea to live site: discovery, information architecture, design, content, build, testing, and launch. Most guides stop there. The version below adds the one thing they leave out: a short search and AI-readiness check at each stage, so the site is findable on the day it launches instead of being patched afterward.
This guide is for business owners and marketers who are about to commission or manage a build, and for developers who want the SEO checks in one place. It walks the seven stages in order, then covers launch day and what to check afterward. For the difference between the roles involved, see web design vs web development.
On this page
- What is the web development process?
- The 7 stages at a glance
- Stage 1 and 2: discovery, then information architecture
- Stage 3 and 4: design and content
- Stage 5: build
- Do AI search features need a different build?
- Stage 6 and 7: testing and launch
- Running the web development process with SEO built in
What is the web development process?
The web development process is a repeatable order of work that turns business goals into a live, maintained website. Every credible description agrees on the shape and disagrees on the count. Figma’s web design and development guide lists seven steps: discovery and planning, wireframing and prototyping, UI design, front-end development, back-end development, testing and launch, then maintenance. The University of Chicago’s website development process splits the same work into thirteen steps, from a kickoff meeting to a pre-launch checklist.
The difference is granularity, not substance. This post uses seven stages because seven fits on one page and each one has a clear exit condition. Neither of those two sources mentions search engine optimization as a stage, which is the gap this guide fills. In practice the cost of that gap is predictable: URLs get changed late, content gets written after the template is fixed, and the first SEO review happens when the site is already live.
The 7 stages at a glance
Each stage ends with a deliverable and a gate. The diagram below shows the seven stages with the search and AI-readiness check that belongs to each.
Scroll the diagram sideways to see all of it.
Stage 5 is highlighted because it is where most search and AI visibility problems are created, and it is the stage most clients cannot inspect. The sections below take the stages in order.
Stage 1 and 2: discovery, then information architecture
Discovery decides what the site must achieve, and information architecture decides which pages exist to achieve it. Chicago describes discovery as an audit of the current site, stakeholder interviews, user research and competitor analysis, followed by a sitemap that serves as the blueprint for the site’s content structure.
Add three search tasks to that work.
- Baseline what exists. Export current rankings, organic landing pages and backlinks from Search Console and your analytics before touching anything. On a rebuild, the pages that earn traffic today are the ones you cannot afford to lose. The website redesign SEO guide covers that case in depth, so this post does not repeat it.
- Map keywords to pages, not the reverse. Decide which page answers which query before the sitemap is final. Our keyword research guide and the keyword research checklist show the method. A page with no query behind it is a candidate for removal from the plan, and a query with no page is a gap.
- Fix the URL structure now. URLs are expensive to change later because every change needs a redirect. Pick short, readable, stable paths. If an existing URL works, keep it.
The exit gate for stage 2: a sitemap where every page has a purpose, a target query where one exists, and a URL that will not change.
Stage 3 and 4: design and content
Design and content are separate stages that must not be separated in time. The most common failure in this part of a project is designing templates around placeholder text, then discovering that real content does not fit.
Chicago’s guide calls for wireframes that outline page features, navigation and information hierarchy before visual design, and for a content strategy that assigns who will find, create, migrate and author each piece. It also notes that authoring and migration is time-consuming and should be scheduled accordingly. That scheduling point is the one projects most often ignore.
Search checks for these two stages:
| Stage | Check | Why it matters |
|---|---|---|
| Design | One H1 per page, headings in order | Headings are how both people and crawlers scan a page |
| Design | Links look like links and use descriptive text | Crawlers and assistive tech both rely on link text |
| Design | Reserve space for images, ads and embeds | Layout shift hurts the CLS metric covered below |
| Content | Answer the page’s question in the first paragraph | Passages that stand alone are easier to quote |
| Content | Name the author and cite sources | Trust signals sit on the page, not in the template |
| Content | Write real copy before templates are final | Prevents truncation and broken layouts |
For the design side, web design for SEO goes through layout decisions in detail. For the writing side, the on-page SEO checklist is the working list, and internal linking explains how pages should point at each other.
Stage 5: build
Build is where the site becomes code, and where three decisions decide whether search engines and AI crawlers can read it: how content is delivered, how fast the page becomes usable, and what the crawl rules say.
Delivery. Google processes JavaScript in three phases: crawling, rendering and indexing. Per Google’s JavaScript SEO basics, rendering happens after a page is queued, and Google still recommends server-side or pre-rendering because it is faster for users and crawlers and not all bots can run JavaScript. The same page lists specifics worth putting in a build spec: use <a> elements with an href for links, use the History API rather than URL fragments for client-side routing, and make sure error pages return a real 404 or carry a noindex tag so you do not create soft 404s. The diagram below compares three delivery patterns.
Scroll the diagram sideways to see all of it.
If your developer proposes a client-rendered app for a marketing site, ask what a crawler receives without JavaScript. Our post on JavaScript rendering and AI crawlers explains why this matters beyond Google, and Astro SEO shows one static-first approach.
Speed. Build to the Core Web Vitals thresholds from web.dev: Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint of 200 milliseconds or less, and Cumulative Layout Shift of 0.1 or less, each measured at the 75th percentile of page loads. These are targets for real-user data, so a lab test is a preview, not a pass. The Core Web Vitals explainer covers how to read the field data.
Crawl rules and markup. Set robots.txt deliberately, and add structured data as JSON-LD. Google generally recommends JSON-LD, and its structured data introduction says not to mark up information that is not visible to the user. See JSON-LD for the format and structured data for AI search for which types earn their place.
Do AI search features need a different build?
No, and anyone selling a separate AI build should be asked to show their evidence. Google’s documentation on AI features in Search says there are no additional requirements to appear in AI Overviews or AI Mode, and no special schema.org markup is needed. A page must be indexed and eligible to appear in Search with a snippet.
Other AI products have their own crawlers, and those are controlled by robots.txt. OpenAI’s crawler documentation describes OAI-SearchBot as the crawler for surfacing sites in ChatGPT search, and GPTBot as the one that collects content that may be used for training. Each is a separate setting, so you can allow search visibility and disallow training. OpenAI notes that robots.txt changes can take about 24 hours to be reflected.
That makes the build-stage decision simple to state: serve content in HTML, keep robots.txt intentional, and write pages that stand alone. Details by crawler are in robots.txt for AI crawlers and in the broader technical SEO checklist for AI search.
Stage 6 and 7: testing and launch
Quality assurance finds what the build got wrong, and launch is the one moment when mistakes become public. Chicago describes QA as testing for errors and design flaws across multiple browsers and devices. Add the search tests below.
| Test | Pass condition |
|---|---|
| Rendering | Key text and links appear in view-source, not only after scripts run |
| Core Web Vitals | LCP 2.5 s, INP 200 ms, CLS 0.1 or better on representative pages |
| Staging leftovers | No site-wide noindex and no Disallow: / in the production robots.txt |
| Titles and canonicals | One title, one H1 and one self-referencing canonical per page |
| Structured data | JSON-LD validates and matches visible content |
| URLs | Every changed URL redirects once to its closest equivalent |
Leftover blocks deserve their own line. Google’s site move documentation notes that some owners block all crawling while in development, and treats blocks that were only needed for the migration as something to remove. A site launched with a staging noindex is a known and avoidable failure.
On redirects, the same page recommends server-side permanent redirects from old to new URLs, says 301 and other permanent redirects do not cause a loss in PageRank, advises against chains, and says not to send many old URLs to one unrelated page such as the home page. For the difference between redirect types, see redirects and SEO.
The timeline below shows what happens after you flip the switch, using the times the sources give.
Scroll the diagram sideways to see all of it.
After launch, submit an XML sitemap, watch Search Console for coverage errors, and run a crawl of the live site against your pre-launch list. The full audit order is in the technical SEO audit checklist, and Search Console is where you will see indexing issues first. Launch is not the end of the process: Figma’s last step is maintenance and iteration, tracking metrics and updating the site.
Running the web development process with SEO built in
The cleanest way to run this process is to treat search and AI-readiness as acceptance criteria for each stage, not as a service bought after launch. That means baselines in discovery, a keyword-to-page map in architecture, real copy before templates, server-delivered content in the build, and a gate at launch that checks redirects, crawl rules and speed.
It also changes how you compare quotes. Ask each vendor which of the seven gates they own, what rendering approach they use, and who handles redirects. For what a project should cost to build and run, see how much a website costs. If you want these checks built into the project from the first stage, Ambeltek’s web development service delivers design, static-first development, structured data and crawler access as one scope.
Frequently asked questions
What are the 7 stages of web development?
Discovery and goals, information architecture, design, content, build, quality assurance, and launch with post-launch care. Agencies label and split them differently (Figma lists seven steps, the University of Chicago lists thirteen), but the order is stable: decide what the site is for, structure it, design it, fill it, build it, test it, release it.
Where does SEO fit in the web development process?
At every stage, not at the end. Keywords and page types shape the site structure, URLs are fixed before design, content is written before the build, rendering and speed are decided during the build, and redirects and indexing checks gate the launch. Adding SEO after launch means reworking pages that already exist.
Do I need a different process to show up in AI answers?
No. Google states there are no additional requirements for AI Overviews or AI Mode beyond being indexed and eligible to appear in Search with a snippet. The same build decisions help other AI search crawlers: server-rendered content, open crawl rules for the bots you want, and clear text on the page.
What is the difference between web design and web development?
Design decides how the site looks and how visitors move through it. Development builds it in code and keeps it fast, secure and working. In a project the two overlap heavily, which is why this process treats design and build as neighboring stages with shared checks.
How long should redirects stay in place after a launch?
Google recommends keeping them for as long as possible, generally at least one year, so search signals can transfer to the new URLs. If you changed no URLs, you need no redirects, which is the cheapest outcome.
What should be tested before a website launches?
Function and layout across browsers and devices, performance against the Core Web Vitals thresholds, that important content appears in the HTML a crawler receives, that staging blocks (noindex or robots.txt) are removed, and that every old URL has a one-hop redirect if URLs changed.