The purpose of this doc is to give a 30,000-ft overview of the Mira attribution process for out of home (OOH).

Attribution is the process of performing studies to determine if there is an effect of media on a particular conversion or other KPI, such as revenue. Mira pioneered the attribution process in the OOH space, and has the most robust and trusted platform for OOH attribution.

Exposure

Mira determines what users are exposed to an out of home campaign by leveraging our mobile location dataset. Exposure can be computed for static traditional , digital, or mobility (moving) OOH. Exposure can also be computed for non-standard environments, such as arena sponsorships.

The result of the exposure process is an "exposure file", which contains an advertising ID, timestamp of exposure, and any metadata that may be leveraged in an "exposure side" cut, such as format or vendor.

Control

Because we cannot control who is exposed to OOH once a campaign is live, we must construct our control after-the-fact. The goal is to have a control group that is as similar as possible to the exposed group, except for the fact that users in the control were not exposed. To do this, we build a control that is modeled on the exposed group across a few variables: demographics, fidelity (quality), and geography.

Conversion

The conversion environment is the set of events which we are trying to attribute to OOH exposures. There are a few we can work with, the out-of-the-box options being: web conversion, app actions, footfall, and brand surveys.

Attribution

The control and exposed files are cross-referenced with the conversion events. An exposed user is said to have converted if they have a conversion event within a certain amount of time after being exposed. Any subgroups are broken down and cross-referenced on their own.

Hypothesis Testing

All OOH attribution is based on sampling the full population. In this way it is similar to political polls in that there is a margin of error around the metrics that are observed. Therefore, it is important to quantify how confident we are in the result we observed, and whether we believe it's true of the whole population.

📘

Statistics is Fundamental to OOH Attribution

Because of all OOH attribution is based on sampling a portion of the population, statistical hypothesis testing is fundamental to the process! Therefore, confidence levels must be reported in order for a result to be usable.