Time series and forecasting · insight-0247
Do not answer a price what-if by changing the price input of a demand forecast
- Evidence type
- documented fact
- Difficulty
- hard
- Roles
- Data scientist, Data analyst
- Industries
- aviation, retail, e-commerce
Summary: situation, action, reason, boundary
Situation
A manager asks what bookings would be if fares were cut 5%, and the analyst re-scores the forecasting model with lower fares.
Action
Say that the forecast is predictive, not causal. Answer with an experiment, a quasi-experimental design or a model with an identification strategy for price, and state its assumptions.
Reason
A causal effect requires assumptions about how the intervention changes the system; predictive associations learned under confounded pricing do not provide them.
Boundary
Price, promotion, capacity and marketing what-if questions answered from observational data.
Symptom or risk
Fares were historically set higher when demand was expected to be high, so the model may have learned a positive fare-demand association and predict that cutting fares reduces bookings.
Exceptions
If prices were randomised or varied for reasons unrelated to demand, the model's price response is closer to causal.
Trade-offs
Causal answers take longer and are less precise; deliver a range and the plan to tighten it.
Pitfalls
Reading a negative coefficient as a valid elasticity just because its sign looks right.
Example
The synthetic airline simulator sets fares higher when demand is expected to be high, while true elasticities are negative.
Lessons and concepts
Claims and their sources
Causal effects are defined by contrasts of potential outcomes under interventions and require identifying assumptions such as no unmeasured confounding.