Similar audiences are determined based on user overlap. More specifically, we use the Jaccard Index, which, for two sets, is defined as the size of the intersection divided by the size of the union. For a given audience, we compare it to every other public audience in the same country, estimating the resulting Jaccard Indices using MinHash. Similar audiences are those with the highest Jaccard Index when compared with the original audience. Audiences require a minimum index of 0.1 to be displayed in this section.
Trending audiences are those with the highest percent increase in size over the past month. Audiences can appear in this list for two main reasons: either they’ve seen an influx of visitors or we’ve recently added many new locations to the audience, greatly expanding its scale.
Demographics are determined based on user home locations and census data. Devices which appear frequently (roughly 20-25% of devices in our data sets) are mapped to probable home postal codes based on their activity. The postal codes are then joined with census data to produce a demographic distribution for the device. For audiences, we then aggregate metrics for individual devices to estimate values for the entire audience.
While demographics can be a useful tool for analyzing audiences, care should be taken when interpreting the results. Some audiences naturally skew towards specific demographic groups and these trends may not be apparent from postal code level data alone.
Currently, RainBarrel builds audiences with data from Canada (CA). If there is another country you are interested in, please contact us.
The size of the audience will fluctuate based on a variety of factors. The most common reasons include: changes in foot traffic and addition or removal of locations. Less commonly, variations may be a result of RainBarrel onboarding or removing data partners. Such cases will be shared in our release notes.
A match rate is the percentage of your audiences that was able to be connected to an ad platform’s users. This number will be unique for every platform and fluctuate as the audience and the platform add and remove users. It can also take anywhere from 24-72 hours for a new audience to be matched by an ad platform. Don’t expect a 100% match rate — the industry average is anywhere between 29% and 62%.