The Staircase Effect, Confirmed in Google’s Own Data
The staircase effect predicts that a new authoritative corpus is adopted by search systems in discrete steps — classification (impressions appear at deep positions), compression (positions climb in jumps), then capture (clicks arrive) — rather than as a smooth curve. Google Search Console data for a 75-day-old domain now shows step one and the onset of step two, in Google’s own numbers.
Key Findings
- Category head-term impressions went 0 → 700 → 2,250 → ~3,300/month pace in the domain’s first ~75 days.
- Those impressions sit at positions 64–95: “answer engine optimization” (600+ impressions), “generative engine optimization” (330+), “ai citation optimization” (~190).
- Zero clicks on head terms — exactly as the model expects at classification depth. Impressions precede position; position precedes clicks.
- Meanwhile, exact-match definitional pages on the same domain already rank positions 3–7 with 8–15% CTR — vocabulary captures its SERPs before category terms do.
- Navigational-era baseline for contrast: the domain’s first partial month showed 29% CTR on a handful of brand searches — people who had heard the name.
The prediction
The staircase effect was defined from earlier deployment telemetry: adoption of a new corpus by a search system is stepwise. First the system classifies — it begins showing the domain for the topic space, at deep positions, generating impressions but no clicks. Then position compression — rankings jump in discrete moves rather than drifting. Only then does click capture begin. A domain in step one looks like a failure to anyone reading clicks alone. It is not failing; it is being filed.
The receipts
| Month | Clicks | Impressions | Phase |
|---|---|---|---|
| April (5 days) | 25 | 86 | Navigational era — brand searches, ~29% CTR |
| May | 12 | ~700 | Classification begins — category impressions appear |
| June | 4 | ~2,250 | Classification at full rate |
| July (8 days) | 1 | ~870 (≈ 3,300/mo pace) | Still climbing; compression onset expected next |
| Surface | Example | Impressions | Position | CTR |
|---|---|---|---|---|
| Category head terms | “answer engine optimization” | 600+ | ~80 | 0% |
| Category head terms | “generative engine optimization” | 330+ | ~86 | 0% |
| Definitional pages (coined vocabulary) | glossary entries | lower volume | 3–7 | 8–15% |
The reading
Two thousand impressions per month at position 80 is not noise — it is the search system announcing which shelf it has decided the domain belongs on. The volume of deep impressions measures classification confidence; the position measures compression progress. They move on different clocks. Watching them as separate instruments — rather than staring at a flat click count — is the difference between reading the staircase and mistaking step one for a plateau.
The secondary finding stands on its own: the domain’s invented vocabulary already ranks top-5 and converts while its category terms sit at position 80. Definitional preemption is faster than categorical competition. That is a measurable, reproducible ordering — and it is the subject of this series’ provenance itself.
Terms Demonstrated in This Note
- Staircase effect
- Stepwise adoption of a corpus by a search system: classification (deep impressions), compression (discrete rank jumps), capture (clicks). Adoption is not a smooth curve.
- Position compression
- The compression phase of the staircase: average position improves in jumps as the system re-files the domain upward within an established classification.
Related Field Notes
FN-003: Entry-Page Decentralization: A Server-Side Gauge of Topical Authority · FN-004: Position 2, Zero Clicks: The Absorption Fingerprint · FN-007: Three Verticals, One Curve: The Ingestion Sequence Replicates