...
We first make sure that the panel of mobile devices is a representative sample for determining home location and trips, by observing the mobile signal and rest patterns. Roughly 83k devices observed over the 50 day period qualified for frequency analysis in Los Angeles. From there, we compare the observed mobile data with traffic counts on particular roadways from different DOTs, and live traffic counts. At this stage, we know there is a high level of alignment between the mobile data and the DOT data. Additional information on the methodology and validation can be found here.
Once we are confident in the trip volume, and that we’re able to observe a device over time, we are therefore able to evaluate the frequency of trips over time as devices pass by OOH inventory. Knowing that the panel of devices is representative of the movement in the market, we can then conduct Reach and Frequency analysis based around this. Device-level data allow us understand Reach and Frequency more closely, as we’re able to understand duplication and model at the device and inventory level.
...