Unlocking the Mobile Advertising Opportunity: Getting Attribution Right
Historically, measuring mobile advertising ROI has been problematic. The holy grail for brands is to tie campaign dollars to a specific dollar transaction or call to action, but data standards, technology, and scale still present big challenges for achieving these measures in the mobile ecosystem. Retail and product advertisers’ best option today is to use mobile location data to measure foot traffic. But, while a foot traffic metric improves marketers' mobile measurement capability, it’s important to note that there are different methodologies and data sources behind this type of attribution measurement.
What should marketers be aware of when choosing an attribution solution?
Today, there are two main approaches to measure foot traffic: auction-based and mobile consumer panels. And, while they measure the same thing, their results may vary widely based on the very different methodology that is used. It’s incumbent upon marketers to understand which method is being used, what data is actually being collected, and how accurate that data is.
When a person interacts with an app and an ad request is sent, an app publisher also sends location data. Auction-based measurement providers typically “listen” to these ad requests to identify whether or not a user is at a targeted location when the ad request is sent.
If they are, a visit to that location is counted as “positive” foot traffic. Solution providers then perform a match to see which of these mobile users were exposed to an ad that is part of the campaign they are measuring. If they have been, they attribute that foot traffic to that ad being seen.
With this approach, the number of potential mobile consumers is larger than is available via panels. Thus the potential audience can mirror the specific demographics selected for a campaign. However, not all of the user’s real-world activity is recorded.
This approach only captures location data when a user happens to have an app open, and when that app sends out an ad request. Also, measurement providers do not have a direct relationship with mobile publishers and do not have control over the accuracy of the data being sent in an ad request. They are at the mercy of a publisher’s own accuracy thresholds. Furthermore, the true location of a user can vary significantly from the stated location in the ad request, which means visits to a store might be recorded when, in fact, there was no visit or equally misleading, the reverse.
Consumer Panel Approach
Alternatively, consumer panels for foot traffic work somewhat like the Nielsen Set Top Box panel. With this approach, an attribution measurement provider creates a large opted-in audience—the “panel”—of mobile consumers who agree to have their location tracked continuously. The measurement provider sees who does and does not visit a specific store. Users can then be tracked back to a mobile ad campaign to determine who did or didn’t see a specific ad.
Since data collection is “always on,” data for each user on the panel is more complete and includes both coverage and duration of visit metrics. Additionally, data accuracy is superior and data quality is higher because the measurement provider is able to control exactly how the location data is gathered.
Panels typically have less scale and their demographics may not be representative of the campaign’s target audience. Additionally, certain panels may incentivize panelists, contributing to a potential demographic skewing of data.
How to make sense of these different methodologies and data sources
Here are eight questions to ask a measurement provider:
- Is your attribution data based on ad calls at a point in time (auction), or do you have “always on” location data (panel)?
- If an auction approach is used, do you have a process for filtering the data for accuracy?
- How do you ensure the accuracy of the data being used in attribution calculations?
- How many mobile users are included in the control and exposed groups?
- Is the attribution audience incentivized and is it demographically representative of my campaign’s target audience?
- How transparent is the methodology to report store visitation?
- What are the primary metrics reported in measuring foot traffic lift?
- Can campaigns be optimized for foot traffic while the campaign is running?
The Future of Mobile Attribution
The good news is that both attribution approaches mean better mobile ROI measurement and better understanding of mobile campaign performance. Over time, solutions will continue to improve and we will get closer and closer to actual purchase or call to action measurements. In the meantime, marketers should ask the hard questions so they fully understand the data they’re dealing with today.