Documentation/Testing & Personalization/Testing & Personalization overview

Testing & Personalization overview

Choose between an experiment, a targeted experience, dynamic Rezonate routing or an incrementality test.

Updated 9 August 2026

Choose the workflow that answers the real question

Incendium keeps experiments, tailored experiences and their first-party measurement in one platform, but the workflows answer different questions. Decide what evidence you need before building content.

A/B test

Compare an inserted component with the unchanged control experience. Use this when one on-page change needs a direct variant-versus-control result.

Split test

Distribute visits between complete URLs. Use it when the alternatives are separate pages or substantially different journeys.

Personalization

Deliver a prepared component to qualifying Audiences or Traits. Use it when relevance, rather than a general winner, is the objective.

Rezonate for Meta

Map Meta Campaigns, Ad Sets or Ads to matched landing pages, custom destinations or AI-generated pages using Smart Links.

Incrementality test

Compare a geographic holdout with matched control regions to estimate causal impact beyond attributed conversions.

Separate the hypothesis from the implementation

Write one decision-focused hypothesis before creating a Component or Campaign: identify the population, the proposed change, the primary outcome and the period or sample needed to make a decision. A visible difference is not automatically a valuable difference.

  • Choose a primary Conversion or performance measure before launch.
  • Check that the relevant Location, Interaction and Conversion definitions already work.
  • Avoid overlapping experiments that change the same element for the same people.
  • Use Audiences for behaviour-derived membership and Traits for immediately supplied classifications.
  • Keep the control experience stable while the campaign runs.
  • Review engagement and impacted outcomes, not only Pageviews.
Better tests begin with better measurement

Because Incendium uses the same journey, engagement, Conversion and attribution evidence across optimisation and reporting, a result can be evaluated in the context of the wider customer journey rather than as an isolated click metric.

Use the shared workflow

Confirm measurement

Verify Locations, Interactions and Conversions before exposing visitors to a change.

Create the reusable Component

Build visually or from code when the experience changes part of a page. Split tests use complete URLs instead.

Configure targeting and delivery

Set dates, probability, audience or trait conditions, placement and any lead limit required by the campaign.

Activate deliberately

Preview the experience, confirm the control and make sure another campaign will not compete for the same placement.

Read the appropriate report

Use A/B, Split Test or Personalization Results with the intended date range and enough observations to support the decision.