Define the question
Choose the channel or activity whose causal impact needs to be measured.
TEST DESIGN
Automatic geographic holdout setup
Choose the activity, test period and region, then let Incendium identify matched control regions with similar historic behaviour. The setup keeps the holdout design, alignment and decision metrics clear before launch.
- Automatically matched control regions
- A clear primary metric with supporting diagnostics
- Test and control periods reviewed before launch
METRICS
Choose the metrics you want to measure
Select the primary metric that defines incremental impact, then add supporting metrics that explain the commercial and behavioural drivers behind any change.
- Revenue per session recommended as the balanced revenue measure
- Conversion rate, effectiveness, Average order value or Total revenue available as primary measures
- Up to six supporting metrics for engagement, acquisition and checkout behaviour
LAUNCH
Launch your test & finally get clarity
Run a controlled test that shows whether marketing created additional revenue, conversions or customer value beyond what would have happened anyway. The result gives your team evidence to scale effective activity, cut wasted spend and defend budget decisions.
- Measure causal lift instead of relying only on attributed conversions
- Separate genuine incremental growth from existing demand
- Scale, reduce or redirect spend with stronger evidence
- Build a repeatable record for future budget decisions
Match the control
Select a test region and review how closely the proposed control regions behaved beforehand.
Run the holdout
Pause or reduce the selected activity in the test region for the complete test period.
Read the lift
Compare the observed result with the control-based expectation and use the supporting metrics to interpret why it changed.
FAQ
Common Questions
How is incrementality different from attribution?
Attribution distributes credit across observed customer journeys. Incrementality estimates how much additional outcome the marketing activity caused. Used together, they show both how customers converted and whether the activity changed the result.
Does Incendium switch campaigns off automatically?
No. Your team makes the planned campaign change in the chosen test region. Incendium provides the test design, control comparison and result measurement.
Why does control-region alignment matter?
The control group represents what may have happened in the test region without the holdout. Similar pre-test behaviour creates a more credible baseline for estimating lift.