SEO Checklist for Developers: What Actually Matters

By Marko Denic
Contents
  1. 1. Make every page crawlable
  2. 2. Make sure Google can index it
  3. 3. Write the page that deserves to rank
  4. 4. Help Google understand it and searchers click it
  5. 5. Make it fast and stable
  6. 6. Earn links instead of building them
  7. What you can skip
  8. 7. Measure the right numbers
  9. Quick recap
  10. FAQ

Most SEO advice is written for marketers. As a developer you control the part that decides whether Google can use your pages at all: URLs, status codes, links, rendering, speed. And you can write content that answers questions better than a generic listicle.

This is the checklist I’d give any developer launching or fixing a site. It follows Google’s own documentation, skips the folklore, and tells you how to check each item.

How a page gets from your server to a click

Each step has to pass before the next one matters. The red marks show what Search Console reports when a step fails.

Work through it in that order. A brilliant article doesn’t rank if Google never crawls it, and a perfect title doesn’t help a page that isn’t indexed.

Download the printable checklist (PDF, 1 page)

1. Make every page crawlable

  • Use real links. Google follows <a href="/page/">. It doesn’t click buttons or follow onclick navigation. If a page is only reachable through JavaScript events, assume Google won’t find it.
  • Link to every page from somewhere. A page that no other page links to (an orphan) gets crawled rarely, if ever. Every post should be linked from a hub, a category page or related posts.
  • Keep the XML sitemap clean. List only URLs that return 200, are canonical and indexable. No redirects, no 404s, no noindex pages. Google ignores priority and changefreq, but uses lastmod if it’s accurate.
  • Don’t block what you want indexed in robots.txt, including the CSS and JS a page needs to render.
  • Answer quickly and reliably. Frequent 5xx errors or timeouts make Google slow its crawling down.

Check it: Search Console → Pages. “Discovered – currently not indexed” means Google knows the URL but hasn’t bothered to crawl it. On a small site that usually means weak internal links, or that Google doesn’t expect much from the page.

2. Make sure Google can index it

  • Return the right status code. 200 for real pages, 301 for moved pages (straight to the final URL, no chains), 404 or 410 for pages that are gone. Every code is explained in my HTTP status codes list. An empty “not found” page that returns 200 becomes a soft 404. On Apache, set up 301 redirects in .htaccess.
  • One URL per page. Pick www or non-www, and with or without a trailing slash, then redirect the other variants. Add a self-referencing <link rel="canonical">.
  • Use noindex on purpose. Thin archives, internal search results and thank-you pages can be noindex. Your articles shouldn’t be. A noindex left over from staging is one of the most common reasons a site disappears.
  • Don’t publish near-duplicates. If two pages cover the same thing, merge them and 301 one into the other. If you serve Markdown copies to AI agents, give each one a canonical Link header that points to the HTML page.
  • Render the content in the HTML. Google renders JavaScript, but sometimes later and not always completely. For content pages, server-side rendering or static generation is the safe choice.

Check it: URL Inspection in Search Console shows whether a URL is indexed, which canonical Google picked, and when it was last crawled. “Crawled – currently not indexed” means Google fetched the page and decided it wasn’t worth keeping. That’s a quality signal, not a technical bug.

3. Write the page that deserves to rank

Since the helpful content changes, this is where most sites win or lose. Google’s core updates reassess how helpful the whole site is, so a pile of weak pages can drag down the good ones.

  • Answer one search intent per page, fully. The reader shouldn’t have to go back to Google.
  • Add what only you can add: first-hand experience, working code, real examples, benchmarks, screenshots. A rewrite of the top 10 results gives Google nothing new.
  • Show who wrote it and why they know the topic: a byline, an author page, links to your work.
  • Keep it current. Update outdated versions, APIs and screenshots. Changing the date without changing the content doesn’t count.
  • Prune or improve weak pages. Old posts with no traffic, no links and no purpose can be merged, rewritten or removed. Google’s advice is to improve first and delete as a last resort.

To find what to write about, start with Search Console: queries where you already get impressions at positions 8–20 are the cheapest wins. More in how to do keyword research and how to write long-tail keyword articles.

4. Help Google understand it and searchers click it

  • Title tag: lead with the main topic, be specific, and stay around 50–60 characters. Google rewrites titles it finds too long, vague or stuffed with keywords.
  • Meta description: not a ranking factor, but it often becomes the snippet. Give a reason to click: a tool, examples, a download.
  • One <h1> that matches the title, then a logical <h2>/<h3> structure.
  • Internal links with descriptive anchors. “CSS button generator” tells Google what the target is about; “click here” doesn’t.
  • Structured data (JSON-LD) such as Article, BreadcrumbList, or SoftwareApplication for tools. It makes a page eligible for rich results; it isn’t a ranking boost.
  • Images: descriptive alt text, sensible file names, and modern formats.

5. Make it fast and stable

Page experience works more like a tie-breaker than a magic lever, but slow pages also lose readers, which hurts everything else. Google measures Core Web Vitals on real visitors:

Core Web Vitals thresholds

Measured on real visitors, at the 75th percentile of page loads.

LCP · Largest Contentful Paintloading
Good ≤ 2.5 s
≤ 4 s
Poor
INP · Interaction to Next Paintresponsiveness
Good ≤ 200 ms
≤ 500 ms
Poor
CLS · Cumulative Layout Shiftvisual stability
Good ≤ 0.1
≤ 0.25
Poor
  • LCP: serve the main image or heading fast. Preload the hero image, don’t lazy-load it, and keep the server response quick.
  • INP: keep the main thread free. Ship less JavaScript and split long tasks.
  • CLS: reserve space for images, embeds and ads with width/height or aspect-ratio, and load fonts carefully.
  • Mobile first: Google indexes the mobile version, so it must have the same content as desktop.
  • HTTPS everywhere, with HTTP redirected.

Practical fixes in web performance.

Links from other sites still matter, but the ones that count are the ones people give you because the page is useful. For developers that’s often a free tool, a cheat sheet, an open-source project or an in-depth tutorial others want to reference. Buying links, link exchanges, and submitting your site to directories can break Google’s spam policies. At best they do nothing.

What you can skip

Where to spend your time

Based on Google’s own documentation, not SEO folklore.

Worth it

  • Helpful content with first-hand experience
  • Crawlable links and a clean sitemap
  • Correct status codes and canonicals
  • Titles that match search intent
  • Internal links with descriptive anchors
  • Fast, stable pages
  • Links you earn by being useful

Skip it

  • The meta keywords tag
  • Keyword density targets
  • A magic word count
  • New dates without new content
  • Submitting to web directories
  • Buying links
  • Schema markup as a ranking hack

7. Measure the right numbers

  • Watch clicks first. Impressions and average position are easily distorted. In September 2025, impressions dropped and average position “improved” for most sites overnight because Google stopped counting most rank-tracker traffic. Nothing about real rankings changed.
  • Compare like with like: the last 3 months vs the same 3 months last year, so seasonality doesn’t fool you.
  • Line drops up with Google’s updates. The Search Status Dashboard lists every confirmed ranking update. Wait at least a week after one finishes before you judge it.
  • Export your data. Search Console keeps only 16 months.

Quick recap

  1. Make every page reachable through real links and a clean sitemap.
  2. Return correct status codes, one canonical URL, and noindex only on purpose.
  3. Write the most useful page for one search intent, and keep it current.
  4. Write specific titles and descriptions, and link internally with descriptive anchors.
  5. Keep LCP, INP and CLS in the green.
  6. Earn links with things worth linking to.
  7. Track clicks, compare year over year, and check Google’s update dates before you panic.

Want it on your desk? Download the one-page checklist (PDF).

FAQ

How long does SEO take to work?

Technical fixes such as indexing problems or redirects can show up within days of Google recrawling the page. Content improvements usually take weeks to months, and recovering from a core update often waits until the next one.

Do I need an SEO plugin?

No. Everything a plugin does (titles, descriptions, canonicals, sitemap, structured data) can be done in your templates. A plugin is just a convenient way to do it without code.

Is JavaScript bad for SEO?

No, but content and links should be in the HTML Google receives. Client-only rendering can delay indexing and hide links. Use server-side rendering or static generation for pages you want to rank.