# Technical SEO, crawlers and pixels checklist

Free checklist from Plain Talk Developers · plaintalk.dev/skills/

Everything a marketing site needs so search engines, AI crawlers and ad
platforms can all read it correctly.

**Start here, before you create anything: ask what already exists.** Creating
a second analytics property next to a live one loses the history and reports
on nothing. Reuse the accounts the business already owns.

---

## Ask first

- [ ] Is there an existing Google Analytics property? Get access, do not make
      a new one.
- [ ] Is there a Google Ads account, and is a conversion already defined?
- [ ] Is there a Meta pixel? A Search Console property? A Google Business
      Profile, and who controls it?
- [ ] Is anyone currently running ads? If so, what is the landing page and
      what counts as a conversion today?

---

## robots.txt

- [ ] Exists at `/robots.txt` and returns 200.
- [ ] Does **not** block everything. A leftover staging `Disallow: /` is the
      most expensive two-character mistake in launches.
- [ ] Names the sitemap:

```
User-agent: *
Allow: /

Sitemap: https://example.com/sitemap.xml
```

- [ ] Blocks nothing you want indexed. Check any `Disallow` line against a
      real URL.

## llms.txt

Increasingly how AI search finds and summarizes a site. Cheap to add, and
most competitors do not have one.

- [ ] Exists at `/llms.txt`.
- [ ] Says what the business does, in plain sentences, not marketing copy.
- [ ] Lists the service pages and the service areas with real URLs.
- [ ] Includes how to get in touch.
- [ ] Contains nothing you would not say on the site itself.

## Sitemap

- [ ] Exists, returns 200, and is valid XML.
- [ ] Contains every page you want indexed and **nothing else**. No
      thank-you pages, no admin, no staging.
- [ ] URLs are absolute and match your canonical form exactly, including the
      trailing slash convention.
- [ ] Submitted in Search Console.

## Canonicals and indexing

- [ ] Every page has a canonical tag pointing at itself in its canonical
      form.
- [ ] Trailing slash is consistent sitewide. Pick one and hold it.
- [ ] No page reachable at two URLs without a canonical resolving them.
- [ ] `noindex` is gone from production.
- [ ] Old URLs 301 to their new equivalents if this replaced a site. Map
      every one before cutover, not after.

## Structured data

- [ ] `LocalBusiness` or the more specific type sitewide, with real NAP:
      name, address, phone, matching the Google Business Profile **exactly**.
- [ ] `Service` on service pages.
- [ ] `FAQPage` where there are genuine questions on the page. Not invented
      ones.
- [ ] `BreadcrumbList` on nested pages.
- [ ] Validated. Broken schema is worse than none.

## Metadata

- [ ] Unique title on every page, front-loaded with the thing it is about.
- [ ] Unique meta description on every page.
- [ ] Open Graph title, description and image, so shared links do not look
      broken.
- [ ] One `h1` per page, and it matches what the page is actually about.
- [ ] Every meaningful image has alt text. Decorative images have empty alt.

---

## Tracking, the part that actually matters

A pixel that loads is not a pixel that measures. What matters is whether the
**conversion** fires.

- [ ] Analytics installed and reporting real sessions.
- [ ] The conversion event is defined and fires on the real completion page.
- [ ] You have seen it fire. In real time, with your own test submission, on
      production.
- [ ] Ad platform pixels present on the pages the ads point at.
- [ ] The pixel is not blocked by a consent banner that nobody configured.

**Test it end to end.** Submit a real lead on production and watch the event
arrive. Ad platforms optimize against the signal you send them, so a
conversion that never fires does not just under-report, it actively trains
the targeting toward the wrong people.

---

## Before you call it done

- [ ] Fetch the live site and confirm it is indexable.
- [ ] Rich results test passes on a service page and a location page.
- [ ] Sitemap submitted, no coverage errors after a week.
- [ ] One real conversion recorded, by you, deliberately.

---

**Want the reasoning?** The Crawlers and Pixels skill covers the setup and
the failure modes, and Content Architecture covers URL structure and service
area pages without tripping near-duplicate filtering.

plaintalk.dev/skills/