Causal answers to business questions when a clean A/B test is not an option, including new market launches, offline media, campaign restructures, and heavy-up periods all measured with statistical rigor.
How does January Digital approach quasi-experimental analysis?
The question is not whether your media is working. It is whether you can prove it. When a traditional A/B test is not possible, quasi-experimental methods let us measure true causal lift anyway, using the same statistical rigor without requiring a controlled lab environment.
We use three approaches depending on the business question: Bayesian Interrupted Time Series for before and after media launches, Difference-in-Differences for comparing groups across a change, and Synthetic Controls for building a statistical twin when no natural control exists.
What kinds of questions can quasi-experimental analysis answer?
This method is built for the measurement problems that fall through the cracks of standard attribution. Did a campaign restructure actually improve performance or did the market improve on its own? Did OOH or CTV drive in-store traffic when there is no click trail? Is media driving lift in a new market or is there just organic demand? Did a seasonal heavy-up drive incremental revenue or did it ride Q4 demand that was already coming?
These are the questions that matter most to leadership and the ones most tools cannot answer cleanly.
How does the process work?
We define the business question and identify the best method, build 2+ years of historical baseline data, define the intervention point such as a launch date or campaign change, model what would have happened without the media, and measure the lift as actual results minus the counterfactual with statistical confidence. The output is a causal estimate of media impact that can be used to justify investment, inform future planning, or make the case for scaling.
When does quasi-experimental analysis make the most sense?
It tends to make the most sense when a brand has run media with no click trail, when a structural change has been made and the team needs to prove it worked, when entering a new market and needing to separate media-driven lift from organic demand, or when planning a spend scale-up and wanting causal validation before committing.
What results has January Digital driven for clients through quasi-experimental analysis?
For Tanger, quasi-experimental analysis attributed $33.4M in estimated incremental sales from a $600K Q4 digital investment, with post-media traffic acceleration of 57% demonstrating that spend built lasting awareness beyond the campaign window. For Carhartt, Google PMAX consolidation was attributed to 119,637 incremental store visits. For a leading retail property company, the analysis validated a strategic media pivot that resulted in the immediate reallocation of all traffic campaign budget into higher-performing ad recall campaigns.