Skip to main content

FIELD NOTE · 12-WEEK ENGAGEMENT

Rebuilding affiliate hygiene for a LATAM marketplace.

A 12-week rebuild of postbacks, partner tiers and bonus rules — and what changed when last-touch stopped being the only ruler.

+22%

incremental revenue

−71%

fraud / bonus claims

−14

active partners

−9%

blended CAC

The client is a Tier-1 LATAM marketplace, two-sided, ~12M MAU, with a four-year-old affiliate program operated through a third-party network. By the time we were called in, gross affiliate revenue had grown 40% year-over-year for two consecutive years — but blended CAC had risen, organic search demand had flattened, and the head of growth could no longer reconcile the numbers between the affiliate dashboard, GA4 and the data warehouse.

What we found

  1. 01

    Last-touch attribution was rewarding the partner that fired the final cookie, regardless of whether the user was already mid-funnel from paid search the same day.

  2. 02

    Three partners had quietly built browser-extension cashback tools that activated on checkout pages, claiming credit from sessions they had not originated.

  3. 03

    Bonus tiers stacked: a partner who hit 'gold' could double-claim on referrals from a sub-tier, creating an incentive to onboard fake sub-affiliates.

  4. 04

    The third-party network's postbacks fired client-side. ITP and ad-blockers caused 18% of legitimate conversions to register late or not at all.

What we rebuilt

  1. 01

    Server-side postbacks via a thin internal collector. Every partner submission resolves a deterministic ID against the warehouse before crediting.

  2. 02

    An incrementality test rotation — 50/50 holdouts on a rolling 4-week basis — to separate compounding partners from those harvesting demand.

  3. 03

    Tier rules rewritten: payout is a function of incremental contribution, not gross referrals. Bonuses are flat at ceiling, not stacked.

  4. 04

    A weekly partner letter (not a dashboard) that explains the reading. Three partners renegotiated; two left voluntarily.

What changed

Twelve weeks in, the program is 14 partners smaller. Gross affiliate revenue is down 8%. Incremental revenue — the number that actually maps to growth — is up 22%. Blended CAC dropped 9%. The partner letter has become the single artefact the head of growth shares with the board.

What we are taking forward

Affiliate is not broken; the ruler used to measure it usually is. A program is healthy when shrinking it makes the rest of the funnel breathe — and that read is only available once tracking, attribution and the incentive structure are all looking at the same events from the same source.

NEXT STEP

Let's design your next season.

Tell us what you are building. We reply within one business day with an honest read on where the real levers are.

Start a conversation