“We need twelve percent growth.”
- What it means
- Commercial target
- What RabbitHawk checks
- Keep it as a target, not demand
RabbitHawk Forecast Interventions
The forecast is saved. Then real life changes.
A supplier runs late. A promotion moves. Roadworks block a store entrance. RabbitHawk captures what changed, tests which forecasts and plans it could affect and shows planners what to review before anything official changes.
Modelled planning walkthrough
See what happens when roadworks threaten store demand
RabbitHawk captures the notice, asks for missing facts, tests whether the forecast and a still-open decision should change, and keeps the saved forecast as the reference until a planner approves an update. Need the connected action plan? See Forecasting & Optimization.
The language layer organizes the evidence. Versioned numerical models calculate the forecast. A planner approves any official change.
Why this is hard today
The number changes. The reason gets lost.
A store, region and planner may each know something useful. But when they edit the number one after another, the final forecast no longer shows what changed, why or who approved it.
- 01
The reason disappears
A useful supplier email or customer conversation becomes a number with no reusable evidence attached.
- 02
Changes overwrite each other
The next edit replaces the last, so nobody can see which evidence moved the forecast.
- 03
A target becomes demand
What the business wants to sell is mistaken for what customers are likely to buy.
- 04
The forecast changed. The plan did not.
Pack sizes, timing, stock and other limits can leave the same order or transfer in place. Check before triggering another replan.
Not every message means more or less demand
Change the right thing
A target belongs with goals. A production limit belongs with supply. A customer update may change demand. RabbitHawk sorts that out before any forecast is recalculated.
“We have too much stock.”
- What it means
- Inventory state
- What RabbitHawk checks
- Test an inventory decision
“Production can make only 5,000.”
- What it means
- Supply constraint
- What RabbitHawk checks
- Use it as a supply limit
“The promotion has changed.”
- What it means
- Potential known driver
- What RabbitHawk checks
- Check whether it is already in the forecast
“A distributor heard orders may fall.”
- What it means
- Weak demand evidence
- What RabbitHawk checks
- Watch for stronger evidence
“The buyer disclosed cash pressure.”
- What it means
- New demand evidence
- What RabbitHawk checks
- Use it as demand evidence
How it works
Ask people what changed, not what number to type
Capture what happened
Turn an email, notice or conversation into evidence with a source, time and affected business scope.
- Who said it and when
- Products, locations and weeks involved
- What is known and still unknown
Ask only useful questions
Ask for facts that could change the assessment, not a percentage override or another forecast number.
- Check what is already in the saved forecast
- Separate facts, assumptions and unknowns
- Estimate the likely size and timing of any effect
Update only what is affected
People contribute evidence and correct assumptions. RabbitHawk's probabilistic model computes the forecast for the affected scope.
- Affected scope only
- Original forecast stays visible
- Still-open decisions are checked again
Learn when outcomes arrive
Compare every saved forecast with what happened and improve the process without ranking people.
- Each saved forecast version scored separately
- MAE, MASE and interval coverage
- No-change recommendations measured too
- Model updates remain governed
Keep the record
Learn from the change, not by blaming the person
Keep the saved forecast, the evidence, each review and the eventual outcome. Compare every saved forecast version against the same outcome to learn which kinds of evidence deserve attention.
- 01
Baseline
Exact incumbent distribution and cutoff
- 02
RabbitHawk
First context-conditioned forecast
- 03
Human
Reviewed belief and reforecast
- 04
Approved
Governed forecast version
- 05
Published
Official forecast delta
- 06
Actual
Outcome and scoring evidence
Evaluate the intervention, not the person
RabbitHawk looks for repeatable patterns across comparable cases. A miss on one case does not prove bias, gaming or poor judgment.
Learning proposals
- Calibration
- Source reliability
- Timing
- Magnitude
- Duration
- Question usefulness
Evaluate before you expand
Keep the planning system you already use
- 01
Choose a representative business subset
Select the products, locations, forecast cycles and decisions that make the evaluation meaningful.
- 02
Connect only the data it needs
Use the history, actuals, product and location structure, and saved forecast versions needed for a short audit or shadow evaluation.
- 03
Keep official planning unchanged
Compare the results and approval flow before deciding whether to connect more of the business.
A practical first evaluation
Start with a short, scoped evaluation
Choose a representative subset of products, locations and forecast cycles. Run it for a defined period in audit or shadow mode while the official plan stays unchanged.
Evaluate a focused part of your business
Choose a representative subset of products, locations and forecast cycles. Run it for a defined period before deciding whether to connect more.
Scope a short evaluation