Runbear Signup Funnel & Landing Conversion Analysis — 2026-06-03

Executive summary

• PostgreSQL and PostHog query access are working. GitHub CLI is authenticated and the plugbear repo is readable.
• DB public."User"."createdAt" is usable as signup truth: 9,483 total users, 9,295 non-Runbear-email users, from 2023-11-14 to 2026-06-03.
• PostHog project currently does not contain production runbear.io / www.runbear.io $pageview data for the last 365 days. Production landing conversion rate is therefore blocked.
• PostHog signup-like events are not reconciled with DB signups: last 365 days DB has 3,413 external signups, while PostHog has only 182 create_user events and 42 user_signup events.
• Landing-change correlation can only be done against DB signup counts, not true landing conversion. Signup velocity declined from late 2025 through Apr 2026 and modestly recovered after the May 12 V3 redesign, but causality is not provable without production landing denominator.

Access status

Signup truth from DB

Source: public."User", columns include id, email, createdAt, updatedAt, profile fields. Default exclusion: email ILIKE '%runbear%'.

Monthly external signups:

PostHog landing coverage

Production landing coverage check:

SELECT properties.$host AS host, count() AS pageviews, uniq(distinct_id) AS visitors
FROM events
WHERE timestamp >= now() - INTERVAL 365 DAY
  AND event = '$pageview'
  AND properties.$host IN ('runbear.io','www.runbear.io')
GROUP BY host;

Result: no rows.

Host categories over last 365d $pageview events:

Implication: this PostHog project is dominated by personal/staging/preview/local traffic. It is not safe to compute production landing-to-signup conversion from this project.

Signup event reconciliation

Candidate signup-like events over last 365d:

DB external signups over same period: 3,413. Therefore PostHog signup events undercount DB signups by an order of magnitude and should not be used as signup truth.

Observed journey patterns — limited to non-production analytics

Because production landing traffic is missing, these are only staging/preview/development observations.

Converted users, defined in PostHog as distinct_ids with create_user or user_signup, commonly show auth/app paths:

Non-converted users commonly show landing/pricing/auth entry but do not progress to signup-like events:

Bot/noise estimate

Only non-production landing-like hosts were measurable. Signals: pageview-only single-event visitors, low-event visitors, and high-pageview visitors.

Suggested ranges for this PostHog project's landing-like traffic, not production:

Do not apply these percentages directly to production until production analytics is fixed.

Landing change timeline and DB-signup correlation

Verified git commits:

Pre/post DB external signup velocity:

Interpretation:

• DB signup velocity was already declining before several changes; the post-change decreases around Dec/Feb/Mar should not be interpreted as caused by landing changes without traffic/channel denominator.
• V3 is the only checked change with a small positive 14d and observed post30 signup/day movement, but the evidence is still weak because production landing traffic is missing.

Prioritized experiments

ICE score = Impact × Confidence × Ease / 10, rough directional score.

Blockers / next data needed

1. Production landing analytics is missing from the accessible PostHog project. Need to confirm whether production uses another PostHog project, GA, Vercel Analytics, Segment, or missing instrumentation.
2. Signup identity join is unresolved. Need canonical event or mapping from anonymous visitor/session to DB User id/email after auth.
3. Current PostHog data is mostly internal/dev/staging, so journey and bot estimates are diagnostic only.

Queries used

Key query patterns:

-- DB signup truth
SELECT count(*) AS total,
       count(*) FILTER (WHERE email ILIKE '%runbear%') AS internal_like_runbear,
       count(*) FILTER (WHERE email NOT ILIKE '%runbear%' OR email IS NULL) AS external
FROM public."User";

-- Production landing coverage
SELECT properties.$host AS host, count() AS pageviews, uniq(distinct_id) AS visitors
FROM events
WHERE timestamp >= now() - INTERVAL 365 DAY
  AND event = '$pageview'
  AND properties.$host IN ('runbear.io','www.runbear.io')
GROUP BY host;

-- Candidate signup events
SELECT event, count() AS events, uniq(distinct_id) AS actors
FROM events
WHERE timestamp >= now() - INTERVAL 365 DAY
  AND (lower(event) LIKE '%sign%' OR lower(event) LIKE '%signup%'
       OR lower(event) LIKE '%register%' OR lower(event) LIKE '%user%'
       OR lower(event) LIKE '%account%' OR lower(event) LIKE '%auth%'
       OR lower(event) LIKE '%login%' OR event = '$identify')
GROUP BY event ORDER BY events DESC;