The Most Common Misuses Of Geo-Data Metrics
‘Store visitation’ as a success metric isn’t much better than ‘clickthroughs’ when it comes to discerning the value of place-based data, says 4INFO’s Chuck Moxley.
But that still doesn’t mean that retailers and their ad agencies understand the best ways to employ geo-data as part of their online-to-offline ad campaigns, says 4INFO CMO Chuck Moxley, who offered his views of the ways brands can go wrong (and right) when it comes to knowing if location ads “worked” or not.
GeoMarketing: What are the most common misuses of geo-data?
Chuck Moxley: The two misuses we see are using a store visit as a conversion metric and using location history to guess who the phone belongs to.
But using store visits to measure campaign success seems to be gaining traction?
Yes, we are seeing more people request it, especially agencies. Hey, “In the Land of the Blind, the one-eyed man is king!” Marketers are so desperate to get a meaningful read of whether their mobile advertising is working, store visits still beat traditional metrics like clicks. There are three key reasons why I think that’s a bad idea.
The first issue is even thinking of it as conversion. I’ve been a marketer for 25 years, and if you asked me when a user “converts” I’d tell you it’s when they make a purchase. After all, if a campaign drives 10,000 visits in stores but sales go down five percent, was that a successful campaign?
Secondly, just because you visit a store doesn’t mean you make a purchase. Take grocery stores, for example. There are more than 30,000 UPC’s in any grocery store. How do I know they bought my product? That forces a ton of guesswork and averaging to guess at an ROI.
Finally, it’s such a flawed idea from a technical standpoint. I can only see you if you are using an ad-supported app while you are in a store. Is the mom with three kids in tow going to stop while shopping at Walmart to play a game of Candy Crush? And location accuracy issues can cause you to guess the wrong store being visited, which throws off the numbers completely.
I’m convinced in a couple of years [that we marketers] are going to look back and laugh about how we used to use store visits to measure mobile ad campaign success! Since many of us are now able to connect mobile impressions with in-store sales transactions, why not use actual in-store sales data to calculate a true, accurate return on your ad spend instead of guessing by seeing when people go in a store?
Why is using location history for audience targeting a misuse of geo-data?
Did you see the recent Mobile Marketing Association report? It talked about using the location history to target people and there was an example in the report about a mom shopping with her kids in Walmart. Her location signal was seen at a park and then at a store that sells kids clothing, at her home, etc. We could do that — anybody who’s doing location could, obviously, do that. All you have to do is match up your points of interest and then, string it together using some machine learning to figure out what kind of patterns do you see and then, assign it.
Our argument against that is you could stream those assumptions together and guess that that device belongs to “a mom with kids.” Or you could look at who the device actually belongs to and know that that household has kids, and what ages are those kids, and then use that for targeting. To me it’s two different misuses of location data.
So how should geo-data be used?
The mobile device is such a personal device. It goes everywhere you go. Marketers have come to regard location-targeting within mobile as the shiny new object. They’re going gaga over it. Our whole point is, use data for targeting and for measurement, and use actual sales information as the conversion.
Where does ‘location’ fit into that equation?
Where location fits is delivering content that’s relevant to the advertiser within the context of where they are and what they’re doing. If people are standing outside of the store, you might want to give them a specific offer. If they’re within a half-a-mile of the store they’re also a good prospect for you. We know they’ve purchased the product that that store sells. Give them directions to your store, and show them where you are in relation to their physical position.
Can you provide any examples from 4INFO’s approach?
We’ve done campaigns like one for a packaged food company where they’ll feature six products available at the nearest Target, which happens to be a half-a-mile from where the person is, and show the actual price on sale this week in that Target store. It’s a very effective use of targeting.
The fundamental question is finding the people buying those products. If it’s Pampers, we’re are going to make sure that they buy baby products, that they have a baby, and they’re buying diapers. We’re not going to target someone who may be a half-a-mile from the store, but who hasn’t bought a diaper in 14 years.
So is your point that there is so much data available to draw an actionable demographic profile, therefore geo-data should be used for a more specific purpose, is that the message?
Exactly. When I’ve seen demos from some of these location targeting guys that create audiences based on where people go, and I’ve always thought that is so cool…if you couldn’t do what we do — which is know who the device belongs to and use the data known about that household for audience targeting, which doesn’t require guesswork.