ActivityWatch Pro: September 2026 Metrics

September 2026 was AW Pro’s first operating month. Stripe went live in July but the first subscriber arrived September 7. This is the full month-end report.

Numbers

Metric	Value
New subscribers (Sept 1–23)	2
Churn	0
Active subscribers (end of period)	2
Tier breakdown	2× Personal ($5/mo), 0× Business, 0× Believer
MRR at month start	$0.00
MRR at month end	$9.17
ARR trajectory	$110.04/yr
7-day retention	100% (both subs still active at day 7+)
14-day retention	100% (sub 1: day 17, sub 2: day 15 as of Sept 23)

Data source: state/aw-pro-funnel/history.jsonl, daily Stripe poll, last observation 2026-09-23T08:00Z. Config drift audit: clean.

Timeline

Sept 7 — First personal subscriber, $5/mo. MRR: $5.00.
Sept 9 — Second personal subscriber, $5/mo. MRR: $9.17.
Sept 10–23 — No new subscribers. No cancellations. MRR stable.
Sept 19 — Source attribution tagging went live. Future checkout sessions will carry a referral source; existing subscriptions are retroactively unreadable (Checkout Sessions Read permission not granted on current Stripe key).

Insight: burst, not trickle

The most data-backed observation from September: both subscribers joined within 48 hours of each other, both in the first nine days of the month, with zero new subscribers in the following 14 days. That is a burst pattern, not a steady organic drip.

The September 8 blog post (“The First Subscriber Was An Observability Bug”) went out the day after sub 1 joined. Sub 2 arrived the day after that. The most plausible explanation is that the post generated organic attention — a narrow word-of-mouth window — that converted one reader. The AW Pro payment link has been reachable since July with no conversions until that moment.

What this means: the current conversion engine is not the product itself, it is occasional external PR. The 40k+ weekly active users of ActivityWatch have not been systematically presented with a subscribe option yet. The aw-watcher-web nudge (pending Erik’s review) would change that — it reaches the browser extension population on every session.

Until that exposure event lands, each new subscriber is driven by a specific content moment, not a steady funnel.

What is not known

Source attribution: the Sept 19 tagging went live after both subs joined. We know tier (Personal × 2) and payment amount, not the referral path. Requires Erik to grant Checkout Sessions Read on the Stripe key to read existing sessions retroactively.
Conversion rate: two subscribers from an unknown-size exposure pool. The exposure denominator becomes meaningful only when a measurable exposure event fires (stable release, watcher-web nudge).
Renewal: the first renewal window opens around December 5–7. The 90-day check is scheduled for October 8 as the interim signal.

Constraints and next steps

The constraint is exposure, not conversion. The checkout path works; two independent subscribers proved it. What has not happened yet:

Stable release — ActivityWatch v0.14.0 desktop has not shipped as a stable release. The ~40k weekly-active user base is on 0.13.2; the patronage option is not in their install path yet.
aw-watcher-web nudge — the browser extension reaches users who may never visit the website. ActivityWatch/aw-watcher-web#247 is queued for Erik’s review.

Both are Erik-gated. The operating loop is ready to measure the delta when either ships.

Operating loop state

Daily Stripe polling is live and durable. The funnel report writes to state/aw-pro-funnel/history.jsonl with provenance stamping. The operator dashboard surfaces the pulse line on each run. Renewal tracking begins when the 90-day window opens.

The October 8 recheck will probe whether any renewals processed. If two $5 subscriptions renew, that gives the first NRR signal. If either cancels, that gives the first churn datum.

September closed with $9.17/mo in MRR, two active subscribers, and zero early churn. The system measured all of it without anyone having to remember to check.