Field notes
Position 40 with clean technicals
A small-business finance data site we run has original data built from hundreds of thousands of loan-level public records, thousands of pages, and technical hygiene we would grade top-decile. It averages around position 40 in search, and about 82% of its pages get zero impressions. This note is about what that combination means.
By Mario Bailey · Published · Updated
We audited our own site expecting to find on-page work left to do. The verdict in our own notes reads: the ranking ceiling is off-site authority, links, not plumbing. That is an uncomfortable verdict, because on-page work is the work we control. And to be clear up front, this page is written mid-climb. The site has not broken through yet. The notes are useful precisely because they are not a victory lap.
Past a hygiene bar, on-page stops being the constraint
The site has the things audits check for: original data no competitor publishes, structured data, honest sourcing, value gates on its programmatic pages, a sitemap with the thin tail pruned out, clean canonicals. It still averages around position 40. The lesson we keep relearning is that technical hygiene is a pass-fail bar, not a ladder. Once a site passes, more hygiene buys nothing, because the engine is no longer ranking the site low for something wrong on its pages. It is ranking it low because not enough of the rest of the web vouches for it yet.
This is worth stating carefully, because most sites sitting at position 40 have not passed that bar, and for them on-page work is still the lever. The order matters: earn the verdict that on-page is done before you accept the verdict that it no longer helps.
How to tell you are link-limited
Three signatures from our own Search Console data made the diagnosis:
- Positions cluster in the 30s and 40s across quality pages. The average sits near 40 and does not respond to on-page improvements. When one weak page ranks poorly, that is a page problem. When the whole site's honest work clusters at the same depth, that is a domain problem.
- The near-misses prove relevance. The site's strongest page type sits at positions 5 to 9: page two, close enough that the engine clearly understands what the pages are for. Meanwhile the demand clusters we most want sit several results-pages deep. Relevance is not the gap. Weight is.
- The tail is dark, and not because it is thin. About 82% of pages get zero impressions. We first took the easy explanation off the table: value gates and a 443-page noindex pass removed the genuinely thin tail. What remained unlit is the authority explanation.
One more honest signal: the site is moving, slowly. In a recent 28-day window it drew about 2,800 impressions and 18 clicks, with clicks going from 3 to 15 across consecutive windows. That is real momentum from a very small base. At position 40, momentum is measured in single digits.
What actually moves it, and what does not
What does not: another technical pass. We did the structural cleanups, slimming the sitemap, gating thin entity pages, fixing dates and schema, and we would do them again, but they were correctness work. They made the site honest. They did not make it ranked. Re-auditing a clean site is the most seductive form of procrastination in this business, because it produces the feeling of progress on exactly the axis that has stopped mattering.
What does: original citable data plus distribution. The site's dataset produced a finding with a roughly 1,400x spread between the best and worst entities on its core risk metric, the kind of number a journalist can quote in a sentence. We built the story around it, with the methodology caveats attached, and a sequenced pitch list of the trade reporters who cover the space. Honest status: at the time of writing, the pitches have not gone out, and the position number has not moved. Our own notes put it plainly: links break position 40. That work is a different discipline, relationships and pitching and patience, and no amount of editing your own site substitutes for it.
So the note ends where the site actually is: on-page done, data asset built, distribution pending, average position still around 40. If a future field note reports the breakthrough, it will cite this one. If it reports that the thesis was wrong, it will cite this one too. That is the deal with writing things down before the ending is known.
Update, July 15: the sequel exists, sooner and harsher than we expected. Four days before this note was published, Google had already ended the site's indexing trial, and our 28-day windows hid it. The forensics are in The discovery window.
The diagnosis discipline Why we publish this
Method: the figures on this page come from our own records and Search Console data, rounded as written. The sites are ours, and we deliberately do not name them. Field notes describe what happened on our network, not a promise of what will happen on yours.