Synthetic control method
The synthetic control method estimates causal impact by constructing a 'synthetic' version of the treated unit — a weighted blend of comparison units that closely matches its pre-intervention behavior. The gap between the real treated outcome and its synthetic counterfactual after the intervention is the estimated effect. It is widely used in geo-experiments where a single market is treated. This page explains the construction and its assumptions.
How the synthetic control is built
Instead of picking one comparison unit, the synthetic control method chooses a weighted combination of many untreated 'donor' units so that, before the intervention, the blend tracks the treated unit's outcome closely. That blend becomes the counterfactual: what the treated unit would plausibly have done absent the intervention.
After the intervention starts, the difference between the actual treated outcome and the synthetic control estimates the effect.
Strengths and cautions
Synthetic control suits situations with a single or few treated units — a campaign in one city, say — where classic experiments are impractical. A good pre-period fit is the core requirement; without it, the counterfactual is untrustworthy.
It shares quasi-experimental caveats: confounding events that hit only the treated unit, or donor pools contaminated by the same intervention, bias the estimate. It is often used alongside difference-in-differences and geo-experiment designs.
- Counterfactual is a weighted blend of donor units
- Requires close pre-intervention fit to be credible
- Fits single-treated-unit geo measurement
How it appears in analytics and logs
A persistent gap between the treated unit and its synthetic control after launch indicates an effect; no gap suggests the intervention did not move the outcome beyond the modeled baseline.
Diagnostic use case
Estimate the effect of a campaign launched in one market or region by modeling what that market would have done without the campaign, using other markets as donors.
What WebmasterID can help detect
WebmasterID's aggregated, per-segment traffic and conversion series can serve as the treated and donor-pool inputs for a synthetic-control measurement.
Common mistakes
- Accepting a synthetic control with a poor pre-period fit.
- Including donors affected by the same intervention.
- Ignoring treated-only shocks that bias the gap.
Privacy and accuracy notes
Synthetic control operates on aggregated unit-level outcomes, not individual data, making it privacy-friendly. This is educational, not legal advice.
Related pages
- Difference-in-differences for measurement
Difference-in-differences (DiD) is a quasi-experimental method that estimates the causal effect of an intervention — like turning a campaign on in some regions — by comparing how a treated group changed against how an untreated control group changed over the same time. By differencing out both pre-existing gaps and shared time trends, DiD isolates the incremental effect. This page explains the method, its key assumption, and where it fits in measurement.
- Geo experiments for measurement
A geo experiment divides geographic regions into a treatment group (which sees a media change) and a control group (which does not), then compares outcomes between them. Because assignment is at the region level rather than the user level, geo experiments measure incremental effect without needing cookies, device IDs, or per-person attribution — making them a privacy-resilient complement to touch-based models.
- Baseline and incremental lift
Every conversion total contains a baseline — what would have happened without the marketing — and an incremental portion driven by it. Incremental lift is that incremental portion: conversions a campaign actually caused, over and above the baseline. Confusing the two leads to crediting marketing for sales it did not cause. This page defines baseline and incremental lift and explains how experiments estimate the split.
- Attribution analytics
Per-segment series for synthetic-control counterfactuals.
Sources and verification notes
- Google Open Source — CausalImpact (Bayesian structural time series)Documents counterfactual/synthetic-control-style impact estimation from time series.
Last reviewed 2026-06-24. Facts are checked against primary/official sources where available; uncertain specifics are marked “Data not yet verified” rather than guessed.