Your site has moved, rankings have fallen, and an audit lists dozens of warnings. Or the homepage appears in Google while half the documentation seems to be missing. A red speed score is easy to notice, but it does not tell you which problem prevents a buyer from arriving.
Google describes search as crawling, indexing and serving results in its guide to how Search works. A page does not automatically pass through all three stages. Use that sequence to decide which audit findings need attention first.
Technical SEO starts with which pages Google has indexed
Start with the pages a buyer needs to find, then compare that list with Google's index. The purpose is to establish whether the site's useful answers can appear in search.
Google's Page indexing report shows how many URLs it has crawled and indexed. The total is a starting point, but the important question concerns particular pages. A report can show indexed URLs while omitting the guide or service page that answers the buyer's search.
Google explicitly advises against expecting every URL to be indexed. Some exclusions are acceptable. This makes a target of getting every URL into Google a poor way to judge an audit. Compare the pages buyers should find with the URLs that do not need to appear in search.
A site: search provides a first look. Google's SEO Starter Guide explains that results pointing to your site establish that it is in the index. That first observation does not settle the status of every important page, so follow it with the report.
For missing documentation, use the same order. Identify which useful docs pages are absent, then examine the reported reasons. Do not start by rewriting the whole documentation site because the overall indexed count looks low. First establish which pages the count represents and which expected pages it leaves out.
For each group of missing URLs, record how far Google has got: discovery, crawling or indexing. That tells the person doing the audit which problem to investigate. It also separates a page Google has never found from one it has read and excluded.
- Name the important pages. Identify the service pages, guides and documentation buyers should reach through search.
- Compare with the report. Read the indexed and excluded URLs against that list to find gaps that matter.
- Keep useful exclusions. Leave acceptable non-indexed URLs out of the repair target, so the audit follows buyer needs.
A few settings decide whether whole sections of a site are seen
A shared setting can affect many URLs, so inspect the domain and indexing instructions before treating each missing page as a separate fault. This is especially useful after a site move.
Open the sitemap and compare its addresses with the live site. Next, inspect the canonical tags that identify each page's preferred address. The browser may show the new site correctly while both still name the previous domain, as the recruiting-research example below demonstrates.
After a migration, the website migration checks belong near the start of the investigation. The move changes addresses and can leave references to the previous site. Reading those references together helps reveal a common cause behind many missing URLs.
Robots rules and noindex instructions also belong in this review. Their presence needs to match the intended treatment of the page. The WordPress noindex guide covers that specific investigation; avoid changing an instruction until its purpose is understood.
The status Google reports matters too. A page it has discovered but not indexed presents a different starting observation from a page it has already crawled and left out. Use the guides to discovered, currently not indexed and crawled, currently not indexed for those respective cases.
Keep the finding attached to the affected pages. An audit should show which setting needs correction and which URLs it affects. That record gives the person making the change a clear target, and it gives the later review something specific to inspect on the live site.
Speed work pays once buyers already land on your pages
When the pages buyers need are already appearing in search, examine what happens after the click. Loading time and the route to an inquiry then become practical questions about the visit.
Speed improvements help the people who reach a page. An indexing repair concerns whether search can bring them there at all. Keep those two jobs visible in the audit, with separate findings for page access and loading time. Then a faster homepage cannot obscure missing search pages.
Once visits arrive, follow the page through to the action it asks the buyer to take. In an August 2026 review, a compliance startup's only form sat at 90% page depth. Average scrolling reached 58.6%. The form's location gave the review another concrete problem alongside loading time.
Those figures describe the page and its observed scrolling. They show why technical and website work need to meet at the visitor's path: the page can load, yet the intended action sits beyond the average scroll. A speed score alone would not describe that problem.
Site size also changes the investigation. A crawl built for a startup's product site is a different scope from a very large catalogue. Before applying the same audit to both, establish which pages it covers. A report cannot explain pages it has not examined.
This is where the starting condition changes the priority. If the important pages are indexed and already receive visits, speed and form access can move forward. If those pages remain absent, investigate why they are missing while preserving any useful visitor improvements already made.
Each deploy and tracking change can undo the fixes
Check the published pages after a release, then check the reports used to judge them. Code that looked correct in a pull request still needs a comparison with what the live site serves.
Include the sitemap, canonical tags and robots rules in the review of a release. Pay particular attention when addresses or shared templates change. Read the deployed settings against the earlier repair: the reviewer should be able to see whether the intended domain and indexing instructions survived deployment.
Analytics need their own review because a change in recorded visits can come from how visits are counted. In October 2026, an engineer found that a UK client's cookie banner hid 70-85% of visitors from analytics. The engineer discovered the missing visits during the first week.
That example changes how a team should read an apparent traffic loss. Before assigning another SEO task, inspect whether the measurement changed at the same time. A report that counts fewer visitors can lead to the wrong content or engineering decision if the counting problem remains unexplained.
Automated activity can distort the other direction. On other projects, browser tests by an agent counted as visits, a bot crawl appeared as growth, and Search Console referenced the wrong domain. Each produced a number that needed investigation before it could guide work.
Keep these checks beside the site repair record. The first asks whether the relevant page and settings are correct after deployment. The second asks whether the report represents real visitors to that site. Passing one does not establish the other.
This is the continuing part of technical SEO. Someone needs to read new releases against previous findings and question unexpected changes in the reports. That work prevents the next decision from resting on an old domain, an overwritten setting or traffic created by a test.
Example: finding why a fast site had no indexed pages
For a recruiting-research firm in August 2026, the same review recorded a faster homepage and found that its pages were absent from Google.
| Check | Recorded finding |
|---|---|
| Homepage load time | Fell from 2.5 seconds to 0.7 seconds |
| Pages appearing in Google | Zero |
| Sitemap references | Pointed to an old domain |
| Canonical tags | Also pointed to the old domain |
The homepage loaded faster, while the domain references explained why search discovery still needed work. The indexing finding determined what had to be addressed before search visitors could benefit from the speed change.
Check the index, then settings, then speed
Inspect key-page coverage
Compare a site: search and the Page indexing report with the pages buyers need, identifying important omissions and keeping acceptable exclusions separate.
Read shared settings
Compare sitemap addresses, canonical tags and robots or noindex rules with the live domain and the intended treatment of each page.
Measure the visit
Check loading speed and the path to the form on pages buyers reach, so repairs address what happens after discovery.
Recheck after releases
Inspect the live settings after deployment and investigate whether analytics count real visitors before using a change in traffic to choose work.