Which initiatives really drove results? With Causal Impact in QuantixAI, you can quantify the effect of a campaign, price change, or product launch - separating signal from noise.
When to use Causal Impact
- Marketing campaigns or promos
- Price or packaging changes
- Product launches or feature releases
- Operational changes (e.g., staffing, logistics)
How it works (high level)
QuantixAI builds a counterfactual - what would have happened without the intervention - using pre-period data and relevant control signals. It then compares the observed post-period to that baseline to estimate lift (absolute and percent), with uncertainty.
Run your first analysis
- Open Causal Impact → choose the target series (e.g., revenue).
- Mark the intervention start date (and end date if finite).
- Add helpful covariates (controls) like traffic, price index, or seasonality proxies.
- Run the model to see the estimated effect and confidence intervals.
Reading the output
- Instantaneous lift: spike right after the intervention.
- Cumulative lift: total incremental impact over the post period.
- Credible interval: the likely range of the true effect; if it includes zero, results are inconclusive.
Best practices
- Pick a clean intervention date and avoid overlapping changes.
- Use stable control series that correlate with the target but aren’t impacted by the intervention.
- Check sensitivity (e.g., alternative windows or control sets).
Make your next decision with evidence. Start your 14-day free trial or contact us for help designing your analysis.