Acquisition is the easy part. Keeping them is the business.

Around two thousand active subscribers is where cohort analysis stops being anecdote and starts being math. Below that, monthly cohorts are too small for a churn difference to be distinguishable from noise, and you are better served by fixing the obvious experience gaps yourself. Most brands we work with sit between five thousand and sixty thousand active subscribers and do fifty thousand to one million a month. If you are smaller but growing quickly, the teardown is still worth it as a blueprint for the volume you are about to hit.

Both, but never in that order. Pouring paid traffic into a subscription that loses a third of every cohort by month two is how brands buy their way into a cash crisis. We fix the offer, onboarding and lifecycle first, then rebuild Meta and Google around payback period and lifetime value instead of first-order ROAS. If you only want media buying and no retention work, we are the wrong agency and we will say so on the first call.

We work day to day inside Recharge, Skio, Stay, Loop, Bold, Shopify native subscriptions and Chargebee, plus Klaviyo, Attentive and Postscript on the messaging side. Headless and custom billing stacks are fine as long as we can get read access to subscription events and charge attempts. The platform matters far less than the data quality behind it. If your subscription events are not being logged cleanly, cleaning that up becomes the first deliverable.

Involuntary churn moves first. Better dunning, card updater logic and retry timing usually recover revenue within two to three weeks because failed payments resolve on a short clock. Voluntary churn is slower by definition: a cohort that enters in March only tells you about month-two retention in April and month-three retention in May. Plan on sixty to ninety days before the retention curve for a full cohort can be compared honestly, and remember that results vary by category and margin structure.

Almost never, and it is usually the reason the problem exists. A deep first-box discount selects for people who wanted a cheap trial, not a habit, and those cohorts churn hardest in month two. Worse, standing discounts destroy contribution margin exactly when you need it to shorten payback. We do use prepay and commitment incentives, because paying for three cycles up front is a fundamentally different economic decision than getting half off one.

We treat involuntary churn as an operations problem with an owner and a number, because for most brands it is eight to eighteen percent of total churn and almost entirely recoverable. That means account updater coverage, retry schedules timed to payday patterns rather than fixed intervals, and dunning messages written as service rather than collections. We also fix the quiet killers: expired tokens, address failures and cards that were never eligible for recurring charges. This is typically the fastest revenue we find in the first month.

For the teardown we need read access to your subscription platform, your ecommerce backend, your email and SMS platform, your analytics and a data export of subscription and charge events covering at least twelve months. For build work we need editor access to the lifecycle platform and the subscription flows, plus a named internal owner who can approve copy and offer changes. If we run paid media we need partner access to your ad accounts, never ownership of them. Everything is covered by a mutual confidentiality agreement before any credential changes hands.

You send the form on this page, we review your numbers and reply within one business day. If the fit looks right we run a forty-five minute call about your cohorts, margin and what you have already tried, then send a scope with a fixed price and a delivery date. Once you sign, access is collected in a single kickoff session and the teardown starts the same week. We take on a limited number of subscription brands per quarter so the diagnostic work never turns into a queue.