PE-301h · Module 3

Forecast Governance

3 min read

Forecast governance is the set of rules that define how the forecast is produced, who can modify it, and how modifications are tracked. Without governance, the forecast becomes a negotiation — reps inflate to look good, managers deflate for sandbagging buffer, and the CRO picks a number based on board expectations. Governance ensures the forecast reflects the data, not the politics.

Do This

  • Separate the data-driven forecast (ensemble method) from management adjustments — both are visible, but the adjustment is clearly labeled
  • Require written justification for any management adjustment that moves the forecast more than 5% from the ensemble
  • Track adjustment accuracy — over time, do management adjustments improve or worsen the forecast? The data answers the question.

Avoid This

  • Let managers override the forecast without documentation — undocumented adjustments are unaccountable adjustments
  • Punish forecast misses so harshly that everyone sandbangs — accuracy improves when honesty is rewarded, not when misses are penalized
  • Change the forecast methodology mid-quarter because the numbers do not look right — commit to the methodology for the full quarter, then adjust for next quarter

The most effective governance model has three layers: the system forecast (data-driven ensemble, no human input), the management forecast (system forecast plus documented adjustments), and the commit forecast (management forecast reviewed and accepted by the CRO). Each layer adds judgment to the data. Each layer's modifications are tracked. Each layer's accuracy is measured independently. Over time, this reveals whether human judgment adds or destroys accuracy.