2026-07-10

TL;DR

• The briefing's ~75 "→OUT" movers look alarming but are normal daily volatility: ~60 keywords fell out of top-100 on 07-10 vs a 45–76 daily norm, and 70 re-entered. Non-event. (confirmed)
• Every project is flat-to-up on 07-10. Avg rank improved in 5 of 6 markets; ranked-keyword count rose in AU, BR, DE, IT and held in AT/UK.
• BR is the real story: +14 ranking keywords in a single day (55→69). Entry is accelerating, not stalling.
• SR1's "200→null" in the briefing is a transient collector fetch timeout — the page is live and healthy and ranks #6 (its best yet). No action.
• The two known 404 profile pages are still 404 and still ranking; 301s remain unactioned since 07-09.

What moved and why

The mass "OUT" exodus — a mirage (confirmed). Across all six markets the briefing lists ~75 keywords going to OUT vs their 7d baseline. The instinct is "algo update or site-wide technical fault." It's neither. Daily transition counts show 45–76 keywords drop out of the top-100 every single day and a similar number re-enter; 07-10 sat at 60 out / 70 in — squarely mid-range, net positive. The list looks big only because (a) the briefing surfaces fallers but not enterers, and (b) 07-09 was a local high in several markets, so the 7d comparison flattered yesterday. Within the churn, most fallers are genuine borderline jitter (e.g. DE "escort" oscillates null↔16↔18; "sex chemnitz" 12↔18↔null week-round). A minority were stable-then-gone (DE "escort in rostock" pinned at 3 for nine days → null; "sex in zwickau" 6–12 → null) — worth noting as individuals, but statistically expected inside a 60-a-day drop-out stream. Verdict: quiet day, no event.

BR acceleration (confirmed). BR ranked keywords: 55 (07-09) → 69 (07-10), avg holding ~11. New market laid down two months ago, now compounding. The watchlist "flag if entry stalls" trigger is nowhere close.

UK city-head slide — paused (likely). The 30-day drift is real: UK "…escorts" head terms averaged 3.46 a month ago vs 4.09 now (~0.6 across 325 kws), matching the rollup's Δ30d −0.6. But week-over-week it's flat (4.06→4.09), and the competitive pressure that drove it is easing — adultwork.com actually receded this week (4.2→4.7). The slide looks like it has stabilised rather than continuing.

Investigations

• Is 07-10 a real drop-out event? Ran per-day ranked/avg trajectories (07-06→07-10) and a daily fell-out/entered transition count. Every market flat-to-up; 07-10 churn normal. Ruled out algo update and collector failure (a failure would cut counts, not raise them). Dead end — and the useful kind.
• SR1 (Sunderland). Briefing showed status 200→null + fetch timeout. Live fetch: page renders fully (66 listings, correct title/H1). Rank climbed 10→7→6 across the three days it's been tracked. The null is a collector-side timeout, not an outage. No action.
• 404 profiles. Both still return 404. "Warragul escorts" (stussy-babe41) holds #10 but has slipped 7→10 over five days — early sign Google is starting to discount it. "escort policoro" (vera-ragazza10) improved 11→7 despite the 404.

Opportunities

• BR momentum is the clearest upside: capturing +14 ranking terms/day. Worth checking which BR clusters are entering (city vs "garotas/acompanhantes" intent) to decide where to push internal links next while Google is receptive.
• Warragul & Policoro are ranking top-10 on dead URLs — fixing the 301s converts fragile 404-rankings into stable ones, likely a quick position gain, before Warragul decays further.

Recommendations

1. 301 the two 404 profiles (stussy-babe41 → Warragul/AU listing; vera-ragazza10 → Policoro/IT listing). Open since 07-09; Warragul's 7→10 slide makes it mildly time-sensitive. Low effort, protects top-10 positions.
2. No action on the mover list. Resist reacting to the "OUT" count — it's baseline noise. This is logged in the watchlist so future briefings don't re-trigger the same false alarm.
3. Keep watching UK head terms for 3–4 more days; if 7d stays flat, close the slide case as monthly drift.

Watchlist

Updated in watchlist.md. Highlights: UK slide downgraded to "paused"; BR reclassified as accelerating; 404s still open (Warragul now the urgent one); SR1 resolved as transient; added a standing "daily churn baseline" note (~60 out / ~60–70 in per day) so mass-OUT lists aren't mistaken for events.