Evaluate past forecast variances to set accurate targets

Analyze past forecast misses to identify systematic over- or under-projection patterns. Use this before locking in next quarter's revenue or operational targets.

You getError Summary Table, systematic Bias Analysis and corrective Rules for Next Cycle

Prompt

Prompt
Act as a quantitative demand planner. I will provide past forecast numbers alongside the actual realized outcomes.

Evaluate the forecast errors using these rules:
- Calculate Mean Absolute Percentage Error (MAPE) and directional bias (systematic over-forecasting vs under-forecasting).
- Categorize the root causes into timing discrepancies, pipeline overestimation, or demand volatility.
- State which specific line items or segments show the lowest forecasting reliability.
- For operational planning purposes only; not formal accounting advice.

Output format:
1. Error Summary Table (Period/Segment, Forecast, Actual, Variance %, Error Type)
2. Systematic Bias Analysis (bulleted findings on recurring prediction flaws)
3. Corrective Rules for Next Cycle (3-4 explicit guardrails to apply to future projections)

Here are my historical forecasts and actual results:

Example output

Example for: e-commerce product line quarterly targets vs actuals

  1. Error Summary Table
Period/SegmentForecastActualVariance %Error Type
Q1 Apparel$450,000$390,000-13.3%Demand shortfall
Q2 Apparel$520,000$465,000-10.6%Over-projection
Q3 Accessories$180,000$215,000+19.4%Under-projection
Q4 Accessories$310,000$285,000-8.1%Timing shift
  1. Systematic Bias Analysis
  • Chronic optimism in Apparel: Apparel was consistently over-projected across consecutive quarters by an average of 12%, driven by expectations of marketing campaign lifts that did not materialize.
  • Negative skew in Accessories: Early-year accessory demand was consistently under-estimated due to ignoring repeat purchases from holiday cohorts.
  • Aggregate MAPE: The overall mean absolute percentage error sits at 12.85%, which is too volatile for low-margin inventory planning.
  1. Corrective Rules for Next Cycle
  • Apply an automatic 10% haircut to marketing-attributed revenue until new acquisition channels prove stable for 60 days.
  • Factor repeat customer repurchase curves into the Accessories baseline instead of relying only on first-touch acquisition.
  • Flag any line item with a variance higher than 10% for mandatory bi-weekly tracking instead of monthly review.
Written by AI from this prompt. Your results will vary.

Paste your past projection figures along with actual results by month or quarter into Claude, ChatGPT, or Gemini. Follow up by asking what error threshold should trigger an automatic mid-quarter forecast adjustment.

Make it repeatable

Keep using this prompt without pasting it in every time.

Keep exploring

More in Data Analysis