Frequently Asked Questions

Why are the number of users so small compared to impressions?

Our analysis is based on unique users, as so is more akin to a reach metric. Furthermore, our location data samples around 10% of the US population, so we don’t capture every exposed user.

Why weren’t my results stat sig? How do I guarantee stat sig?

Because statistical significance is dependent on the magnitude of the true underlying effect, it is not possible to guarantee it. Consider an OOH campaign in which creative was a blank screen - we would not expect this to have any impact on consumer behavior, and the true lift would likely be 0. Thus, no amount of increase in sample size would ever allow us to realize statistical significance, because the effect is not there to observe.In many cases, we observe positive lift, but don’t reach statistical significance. Here, the results are promising and may have reached statistical significance with a larger sample size (more impressions).

How can I improve my results? How can I increase scale?

The best way to increase scale is to measure larger campaigns that have additional geographical and temporal “dispersion”. Because our analysis is based on unique users, adding additional impressions in markets that are not currently included is most likely to increase the unique sample set.

Why did the control group have a lift?

The control group is identical to the exposed group except for the fact that it was not exposed to the OOH campaign. It is entirely possible for variables other than OOH exposure (seasonality, exposure to other media, exogenous events such as product launches) to drive a change in consumer behavior. We’d expect these variables to drive changes in both the control and exposed group, while only the exposed group experiences the incremental impact of OOH exposure.

Overall results were stat sig, but my cuts weren’t. Why?

Cuts are subgroupings of the results, and so by definition have smaller sample sizes which impacts stat sig. Furthermore, stat sig is driven by the magnitude of the lift, which can be higher or lower in a cut result than the overall result

How is exposure linked to online actions, such as visiting a web page?

We leverage device graph partners to tie mobile advertising IDs (MAIDs) to a variety of other identifiers and signals.

What other types of measurement do you offer?

In addition to our attribution product (Incremental Lift Analysis for Web, App & Footfall environments), we offer user-targeted brand awareness surveys