Fractyl
Request a demo
Menu

Use cases

Start with the decision.
Then build the audience.

Fractyl is most valuable when it is tied to a specific media or customer problem, not when it becomes another destination for data.

01

Suppression

The strongest first use case is arithmetic, not prediction.

Exclude current customers, opt-outs, invalid records and bulk-import exclusions from acquisition. It needs no model, applies to the largest media budget line, and produces a result that survives commercial challenge.

Good starting audiences

  • Current-customer exclusion
  • Opted-out and invalid records
  • Bulk-import exclusions

Proof to design upfront

Quantify prospecting spend and impressions removed. Validate the saving against any lost reach or mistaken exclusions.

02

Win-back

A dormant customer is invisible to platform targeting but known to the CRM.

Use lifecycle state and a relevant date to re-engage lapsed, dormant or deferred customers around intake, renewal or repurchase windows. No predictive model is required.

Good starting audiences

  • Lapsed and dormant customers
  • Deferred or withdrawn prospects
  • Past customers entering a new need state

Proof to design upfront

Compare reactivation and incremental value against a holdout or agreed baseline, not cold-prospect conversion rates.

03

High-value lookalikes

The quality of the seed determines the quality of the acquisition signal.

Replace the generic list of everyone who ever bought with customers defined by value, speed to conversion or retention. Where modelled lifetime value is not live, use a transparent observable proxy.

Good starting audiences

  • Highest revenue-to-date customers
  • Best retained customers
  • Priority product tier or tenure

Proof to design upfront

State whether the seed uses modelled value or a proxy, then judge performance on qualified conversion and customer value.

FIND THE FIRST USE CASE

We’ll pressure-test value, feasibility and data readiness.

Talk through the opportunity