growth diagnostic
press enter

CAC & Signups — working diagnosis

Evidence-first review of the CAC and new-customer miss, prepared for the Shane / CTO / growth session. Hypotheses are ranked, not concluded — the next steps are proposals for discussion.

Open questions

    Recommended next steps — for discussion

      These are proposals to pressure-test in the session, not orders. The kill-switch experiment is the fastest way to turn hypothesis #1 from correlation into causation.

      Timeline

      90 days, all signals on one axis. Toggle layers; dashed verticals mark the deliberate changes. Noisy daily series default to 7-day moving averages.

      Funnel — where it breaks

      Healthy window vs decline window (Intelligems sitewide). The break is isolated to one step: engaged visitors stopped progressing to add-to-cart. Everything downstream is intact.

      Exonerated steps — flat between windows

      Landing mix — healthy vs decline

      Testing load

      26 A/B tests overlapped the window. The chart below maps every span; the second chart overlays daily concurrent tests against sitewide CVR.

      Concurrent tests vs sitewide CVR (7-day avg)

      Rise 2

      The highest-traffic landing page — and the page most of the May/June tests ran on. Weekly visitors and CVR below; behavior signals from Clarity at right of the stat row.

      * partial week. Includes rise-2-short and rise-2-founder URL variants.

      Clarity behavior snapshot — last 3 days only (API limit)

      Channel mix

      Weekly spend by channel (top 6 + other) with Meta’s share of total overlaid. The mix shifted deliberately in May — the open question is whether the Google increment is finding new customers or harvesting brand demand.

      Mix shift — healthy vs decline window (share of total spend)