Skip to content
ML Academy

Engineering, production, governance, cost, communication · insight-1303

Price the exit before buying a model or platform

Evidence type
practitioner heuristic
Difficulty
medium
Roles
ML engineer, Data scientist, Applied AI
Industries
cross-industry, banking, saas

Summary: situation, action, reason, boundary

  1. Situation

    A team picks a hosted risk-scoring API after a strong demo and a three-year contract is on the table (hypothetical).

  2. Action

    Before signing, write the exit plan: which inputs, outputs and versions you log, whether scores can be exported or reproduced, who revalidates, person-weeks to switch, and notice and price-change terms. Add that cost to the multi-year comparison and keep a simple challenger.

  3. Reason

    Total cost of ownership includes the cost of leaving, and it is largest when decisions and workflows are coupled to one provider's outputs.

  4. Boundary

    Vendor models, hosted APIs and managed platforms embedded in decisions.

Symptom or risk

Leaving later needs re-labelling, re-validation and re-integration, and a price rise or discontinuation finds the team with no alternative.

Exceptions

Short, reversible pilots can skip a full exit plan.

Trade-offs

Logging and a challenger cost money and effort now, against optionality later.

Pitfalls

Planning the exit around data export alone, while the validation work is the real cost.

Lessons and concepts