Skip to content

Configure Tulo Paywall - Experiments

Target Audience: Users, Developers

Introduction

Tulo Paywall Experiments let you A/B-test different paywalls against each other to find out which performs best. An experiment splits traffic between two or more variant paywalls for the same context, so you can compare conversion rates and other key metrics before committing to a single paywall configuration.

An experiment is assigned its own experiment code that follows the same format as a regular paywall code. Publishers use the experiment code exactly where they would normally use a paywall code — in the JWT pc claim (Standard) or as the paywallCode initialization parameter (Lite). No other integration changes are required.

When a request arrives with an experiment code, the system deterministically resolves which variant paywall to show based on the visitor. The resolved paywall is then rendered normally with its own template, offer, and eligibility evaluation.

How it works

Visitor identity

The Paywall Javascript client automatically generates a unique visitor ID, per visitor on the website that implements the Tulo Paywall. This ID is sent with every paywall request and is used to determine which experiment variant a visitor sees.

No action required

The visitor ID is fully managed by the Javascript client. Publishers do not need to generate or handle it.

Variant assignment

When a paywall request includes an experiment code, the system assigns the visitor to a variant using a deterministic hash of the experiment ID and visitor ID. This means:

  • The same visitor always sees the same variant for a given experiment, ensuring consistent user experience and statistically valid results.
  • Traffic is distributed according to the configured weights (e.g. 50/50 or 70/30).
  • The distribution is approximate and converges towards the configured proportions as traffic volume increases.

Note

Visitors who clear browser data or use multiple devices may be assigned different variants across sessions. At sufficient traffic volumes this has negligible impact on results.

What happens when a code is not recognized

If the provided code does not match an active experiment, we will attempt to fetch a Tulo Paywall with the given code instead. If this still fails, an error response will be shown.

Configure an experiment

Experiments are managed in PAP under Marketing => Tulo Paywall.

  1. Select title. Choose the title the experiment belongs to. All variant paywalls must be active for this title.
  2. Set name. An internal name to identify the experiment (e.g. "Summer campaign A/B test").
  3. Add variants. Select two or more active paywalls as variants.
  4. Set traffic weights. Assign a weight to each variant as positive integers representing it's relative proportion. For example, 70/30 or 50/50.
  5. Save. The experiment is created in a stopped state. Activate it when you are ready to start testing. Note that activated experiments cannot be modified, and if you wish to change/delete an experiment, it must be stopped first.

The experiment is automatically assigned a unique experiment code that you use in place of a paywall code in your integration.

Paywall protection

Paywalls that are part of an active experiment are protected from changes:

  • They cannot be edited, deactivated, or deleted while the experiment is active.
  • The paywall list and edit views display a clear indicator when a paywall is part of an active experiment.

To unlock a paywall, stop the experiment it belongs to first.

Using the experiment code

Use the experiment code in the same way you would use a regular paywall code:

Standard integration — set the experiment code as the pc claim in the JWT:

{
  "pc": "your_experiment_code"
}

Lite integration — pass the experiment code as the paywallCode initialization parameter:

TuloPaywall.init({
  paywallCode: "your_experiment_code",
  // ... other parameters
});

The paywall will be rendered normally — the end user will not know they are part of an experiment.

Measuring results

Experiment results are tracked via Engage BI. Please contact our support for more information, or if you need help with exporting results from your Tulo Paywall Experiments.