A/B test paywalls and offerings in your app with audiences and targeting rules

Your app asks for the current offering, or for the offering of a placement. RevenueDot answers per customer, in this order: an offering you assigned to that one customer through the API, then your live targeting rules from the top, then the project's current offering.

An audience is a set of conditions. Conditions in a group must all be true, and separate groups are alternatives. Preview shows how many of your customers match today.

• Fields: country, platform, app version, SDK version, locale, subscription status (active, trialing, expired, never), active entitlements, total spent, first and last seen, latest product, email, attribution (campaign, media source) and any customAttribute:<key>.
• A rule can have start and end times through the API.
• Rules are checked from the top. Use Move up and Move down to order them. New rules start off.

Your app code stays the same. It reads the offering the way it does today, and RevenueDot decides which one comes back.

let offerings = try await Purchases.shared.offerings()

// The offering your targeting rules chose for this customer.
let paywallOffering = offerings.current

// The offering for a placement you named in a rule.
let onboarding = offerings.currentOffering(forPlacement: "onboarding_end")

An experiment compares two offerings: a control (a) and a treatment (b). Each offering can carry its own paywall.

1. 01

   Create both offerings

   Make the control and the treatment offering in your catalog, with the packages and the paywall each one should show.

2. 02

   Create the experiment

   Open Experiments and select New experiment. Pick both offerings, an optional audience and the share of customers to enroll.

3. 03

   Start it

   Select Start. Matching customers are enrolled the next time the app fetches offerings and always keep the same variant.

4. 04

   Read the results

   Wait for at least 100 customers in each variant, then compare conversions, trials, revenue per customer and the chance the treatment converts better.

5. 05

   Pause or stop

   Pause keeps enrolled customers on their variant and enrolls nobody new. Stop ends the experiment for good.

Enrollment is deterministic: a customer who matches always lands in the same variant. Enrolling sends an EXPERIMENT_ENROLLMENT webhook once per customer and experiment, and enrolled customers' purchase, renewal and cancellation webhooks carry an experiments list, so your analytics tool can split by variant.

• Customers per variant.
• Conversions, which count any purchase or trial after enrolling.
• Trials, revenue and revenue per customer.
• The chance the treatment converts better than the control.

Audiences live at /v2/projects/{project_id}/audiences and targeting rules at /v2/projects/{project_id}/targeting_rules. Experiments live at /v2/projects/{project_id}/experiments, with .../results per variant. Start, pause and stop are actions on the experiment. An experiment compares two offerings, so test several ideas one after the other.