Covers ~Jun 2 to 9. The items worth the room's time, then the numbers and the decisions.
Summary: affiliate review pages convert 43 to 72%; the broker funnel's "2.1%" was a tracking undercount, not a leak (real volume ~270 leads/mo, all reaching the CRM); search traffic has bottomed and is recovering.
The "2.1%" was an analytics undercount (~5x low). D1 (source of truth) shows ~270 leads/mo, all delivered to the CRM. Funnel volume is healthy.
Shipped: a live Leads dashboard in the Cloudflare admin hub. Shows every lead and its CRM-delivery status, with a drift alert.
In progress: pushing paid leads into Google Ads server-side (ad-block-proof) so bidding optimizes on real conversions, not the ad-block-shrunk count.
Also: replaced a broker exit popup that was misconfigured and never ran with a live A/B test.
2. Popup engine fixed
An interactive popup logged 0 clicks on 600+ views: injected popup HTML blocks inline scripts, so every interactive popup was dead. Fixed and A/B-verified.
The "24% error rate" is mostly ad-blockers blocking trackers, not site errors. Leads save server-side regardless.
3. Domain-broker concept suite (concepts.vpn.com)
Four intake concepts on the production design system, including "The Brief" (editorial). Client feedback is in; quick polish and qualifying fields are applied.
Production form upgraded: international phone, floating labels, brand-correct buttons.
4. AI search (GEO) as a growth lever
AI referral traffic converts 4 to 10x organic and is growing fast; ours is ~0.2% today. Entry cost is mostly content we already produce.
The lever is authority plus original data: a genuinely real, dated Speed Lab, which earns the citations competitors currently hold.
5. Search traffic recovering (GSC, through 5 Jun)
Last 4 weeks vs. prior 4: impressions +118%, clicks +24%, average position improved 4.7 (29.0 to 24.3).
CTR more than doubled since December (0.47% to 1.20%).
Commercial pages moving to page 1: free-VPN guide 19.8 to 10.7 (clicks 198 to 414); new pricing pages entered on page 1.
Biggest untapped lever: US (133k impressions, 0.16% CTR, ranks ~27). A cohort sits at positions 11 to 20, one step below page 1.
Open: cause of a mid-April impressions drop (algorithm, seasonal, or technical) needs diagnosis.
By the numbers (itbroke, last 7 days)
Conversion rate 11.4% (vs. ~8.8% 30-day average).
Affiliate out-click: NordVPN 61%, ProtonVPN 72%.
Broker funnel: ~270 leads/mo, all delivered to the CRM.
LCP p75 2.27s; server response 171ms.
Traffic mix: ~69% direct/brand, ~19% organic, ~7 to 9% paid.
Decisions for the room
Value model for paid conversions. Confirm: optimize only for buyers above $50k, report a performance-fee value, and increase it on verification (a verified, signed buyer counts more). This is the server-side Google Ads work, in motion now.
Speed Lab real-data call. Make the top-5 provider tests genuinely real (comped/press accounts, ~$0). Linchpin for the AI-authority play.
Confirm the 50-day focus (fix leaks, capture demand, plant authority) ahead of the in-person visit.
Data note
itbroke and GSC (SerpDelta, through 5 Jun) are folded in. The Google Ads bridge is not yet connected, so the paid-spend deep-dive is pending.