Measuring a planning app without measuring its task content

Disclosure: PlanGentle and XPmetric are both built by Stefan Rosanitsch. This is an internal implementation note, not independent customer research, an endorsement, or an efficacy study.

PlanGentle is a daily planning app. This note describes how its implementation connects a consented visit to fixed account milestones and, where separately permitted, checkout attribution. That is an instrumentation question; it does not show whether the app changes planning outcomes.

Measurement starts after opt-in

PlanGentle's consent manager loads the XPmetric cookie tracker and Google Analytics 4 only after a visitor accepts analytics. Its page-measurement component waits for consent and tracker readiness before checking acquisition pages, /start, or /billing.

On an acquisition page, a signed-out visit can be recorded as a landing candidate and emit the fixed pg_landing_view milestone. The server checks consent and signed-out status, then ties the candidate to XPmetric's first-party visitor cookie. Account measurement can emit fixed milestones such as pg_signup_completed, pg_first_task_saved, and pg_returned_day. Missed or pre-consent actions are not queued for replay.

Keep planning content out of analytics

The tracker records measurement fields such as the page path, supported campaign parameters, referrer origin, and XPmetric visitor and session identifiers. XPmetric derives browser, operating system, and device categories from the request. Google Analytics is a separate consented integration: PlanGentle's privacy notice says it receives selected page views and fixed event names, along with browser and device information and an IP address for processing. Page URLs sent to Google omit query parameters and fragments.

Each milestone name comes from a fixed list. Its event call can carry page and campaign context and XPmetric's opaque visitor and session identifiers, but no task titles, first actions, notes, reflections, email addresses, draft identifiers, sign-in tokens, or raw account IDs. PlanGentle keeps a consented visitor-to-account link on its own server for cohort measurement; the export uses a site-specific keyed hash of the account ID. Withdrawing consent removes the stored visitor-to-account link.

For checkout attribution, the browser supplies XPmetric visitor and session IDs only after analytics consent. The server checks consent again and requires the visitor ID to match the first-party cookie before attaching those IDs to Stripe checkout and subscription metadata. A checkout start is not a payment; payment attribution depends on verified provider records.

What this note does not show

This note reports no traffic, signup, conversion, retention, or time-saved results. PlanGentle has no product efficacy study, so this instrumentation cannot support claims that the app improves productivity, reduces anxiety, or treats ADHD.

The point of the setup is to make a later measurement path explicit while keeping user planning content out of analytics. It does not substitute for a product evaluation.

PlanGentle's public privacy notice explains its current data practices. To see the planning flow described here, try PlanGentle.