Sample Estimate Confidence
This white paper discusses the theory behind UberMedia’s sampling methods to determine if the amount of devices observed is enough for making business decisions.
UberMedia collects location data from a large number of mobile devices via both location SDKs and advertising SDKs. If a mobile device has location services active, the installed apps pass along latitude and longitude of the device. By collecting these location coordinates from each request, along with a unique device identifier and the time of the request, we can study spatial-temporal behavior of the people who use mobile devices.
Analyses of these data are complicated by the fact that data points are only available intermittently. Consequently, we only observe a sample of devices and a sample of their locations. The question often arises, is the sample seen within a retail store or boundary “enough” for making business decisions?
When to Send: When someone asks “is X number of visitors enough to base an analysis on and have it be statistically significant?.
Last Updated: February 2019
Other Supporting Material: none
Insights provided by UM in any study mentioned are aggregated and de-identified, not tied to any single device or individual. UM adheres to GDPR and CCPA and has been certified for privacy compliance by Verasafe, an independent third party.