Client work · Consumer discovery

Atlas

Alias · real name on request

Resolving a structural unit economics risk

A core feature of the product depended on a third-party geolocation and points-of-interest API, licensed on a pay-per-call, pay-per-asset basis. During early-stage beta testing with a limited user cohort, monthly spend on this dependency was already material relative to the size of the user base.

Modeling that cost curve forward against projected user growth made the trajectory clear: left unaddressed, cost-to-serve on this feature would scale linearly, and eventually faster than linearly, with adoption. At meaningful scale, the unit economics would not hold.

This was not yet a live crisis. It was a structural risk identified early enough to be redesigned around, rather than discovered later as a margin problem.

Approach

  1. 01

    Own the data rather than rent it

    I made the decision to reduce structural dependency on the third-party API rather than continue absorbing its cost curve, and led the shift toward a proprietary, first-party data layer built in two parallel tracks. The first was a near-term, operationally-led data acquisition process to populate core markets directly.

  2. 02

    An ingestion pipeline for the long run

    The second track, still in active development, is an automated ingestion pipeline designed to source, clean, and structure the same category of data from a broader range of public sources at lower marginal cost.

  3. 03

    A tiered architecture, not a cutover

    Eliminating the third-party dependency outright would have meant losing the product capability it uniquely offered, so I designed a tiered product architecture around it instead. The proprietary data layer became the default experience for the majority of the user base, effectively decoupling cost-to-serve from user growth.

  4. 04

    The expensive dependency pays for itself

    The third-party integration was preserved and repositioned as a premium capability, monetized directly, so the feature's most expensive dependency now funds itself rather than eroding margin.

Outcomes

  • The redesign converted an unbounded, usage-linked cost center into a cost structure that scales sustainably with the user base.
  • The premium capability was preserved and monetized rather than cut, so the most expensive part of the feature now funds itself instead of eroding margin.
  • It also reflects a broader pattern in how I approach product decisions: catching unit economics risk early, from a small signal, before it becomes a scaling constraint.

More work

All case studies