Open experiment log
The organic growth experiment
We published far too many search pages, learned that ranking for weak queries is not growth, and started removing the machinery that created the problem. This is Sipiteno's public account of the baseline, the decisions, and the results.
Where the experiment started
On , we measured the search performance of the portfolio that includes Sipiteno. The scale looked impressive until we compared it with actual traffic.
| Baseline measure | Observed value |
|---|---|
| Published URLs | 16,691 |
| Organic clicks in the trailing 28 days | 134 |
| Click-through rate | 0.27% |
| URLs ranking in the top 10 | 2,921 |
| Ranked URLs that earned a click | 98 |
| Published URLs that earned no click | 99.4% |
The contradiction changed the strategy. Thousands of pages had achieved the technical outcome they were built for, but almost none had earned attention. Adding more pages would have multiplied the failure mode rather than solving it.
What we stopped doing
Page count as progress
A generated URL is now a liability until it demonstrates demand or carries unique data. Shipping volume is not reported as a win.
Rank without demand
Positions and impressions are diagnostic signals. They do not matter when the underlying query has no useful audience.
Unsupported authority
Fabricated bylines, office presence, and unsourced performance claims are being removed. Trust is harder to rebuild than a page.
What changed on Sipiteno
We removed hidden prerendered text, deleted a fabricated research byline, replaced false local-presence wording with remote-delivery language, consolidated duplicate country-service routes, and stopped advertising pages that did not exist. The work is deliberately subtractive.
The next phase is to publish fewer things that are independently useful: source manifests, downloadable evidence, decision tools, and honest postmortems. Every new indexable page must replace an existing weak page. This log replaces the old Dream list page.
Latest measured state
The current Sipiteno snapshot covers through and was measured on .
| Measure | Observed value |
|---|---|
| Sitemap URLs | 280 |
| Organic clicks | 14 |
| Search impressions | 3,625 |
| URLs earning a click | 10 |
| URLs ranking in the top 10 | 82 |
| Indexed-and-impressed ratio | 88.9% |
This is not a victory lap. Click volume remains small, and most top-ranked URLs still do not earn a click. The useful change is that the measurement now makes the failure visible and prevents us from hiding it behind page count.
How to audit this log
The measurements behind every number on this page are published as a machine-readable manifest. It records the observation window, value, source URL, and measurement time. When the data changes, the manifest and this page change together.
Download the source manifest (JSON)
What we expect to learn
We are testing whether a smaller, more honest site can earn attention from a wider set of pages than a large templated site. Falling impressions can be healthy when low-value queries disappear. The signals that matter are clicked URLs, qualified demand, references from independent sites, and whether a useful artifact is downloaded or cited.
If those signals do not improve, the experiment should say so plainly. The point is not to make Sipiteno look right in retrospect. The point is to create evidence strong enough to change what we do next.