KEVIN OUTCALT: SAP mobile services is a division of SAP that’s a little different. We don’t sell software licenses like virtually all of SAP does. We sell services to the mobile network operators and enterprise customers and primarily those services are messaging services. So if you’re sending a text message from your home to London and it goes through a local operator and has to be transferred to say, Vodafone in the UK, it goes through us. That way the thousand or so other mobile network operators don’t have to have agreements with the other operators.
As we developed, we started to understand that the enterprise customers are in the dark about what mobile user behavior is and where to reach them and with what kind of message. On the other side, we’re aware that mobile network operators have a tremendous amount of data about mobile user behavior. And so the service we developed, which we call SAP Consumer Insights 365, puts those two pieces together.
What kind of data are you collecting?
KO: We’re working with the mobile operators to receive completely privatized, anonymized and aggregated data that provides marketing intelligence for brand agencies and brands that want to understand at a population level what is going on with mobile user behavior, where and at what time. The types of information we collect are things like location, time, clickstream information for websites, age, gender and to some extent home location. All we receive are the first two digits of the zip code so you can’t identify an individual, but you can identify users by city.
With that information, we’re able to put together a subscription to a market intelligence service where advertisers can search by those different data attributes to get more information about what’s going on in the mobile environment at specific places at specific times so they can deliver appropriate ads.
How are you processing the data?
KO: We run it through HANA [High performance ANalytic Appliance in-memory database] which is what makes this possible. The sheer size of the data, where even a medium-sized mobile operator has a couple terabytes a day, up to now has been really hard to process because of the size and the response time needed to run queries. With HANA, where everything’s in-memory, that allows us to be 1,000 times faster than conventional technologies that are used for doing searches. That means you’re getting sub-second response times on your queries instead of minutes or longer. Then we provide a custom portal through the subscription that gives marketers the ability to log in and search based on their subscription level and the interests that they have as a business.
Does SAP have any interest in selling media or creating an ad exchange?
KO: The brands still make their own media decisions. Once they understand the market intelligence, they can map that against their own strategies. We don’t do any ad serving or link to any specific media for purchase.
When will Consumer Insight 365 be available to marketers?
KO: We’re sitting down with the mobile operators to put together the relationship agreements. They’ll receive a revenue share for every subscription that we serve to an end-user. Our plan is roll this out to the brands in late Q3. We’ll be rolling it out by country in the UK, US and Western Europe, but not necessarily in that order.
Guy, how is Kantar Media involved with SAP?
GUY ROLFE: When SAP told us they were looking to create this service, we helped explain to them the important metrics the advertising industry would need. We would like to become a customer of SAP Consumer Insights, but it is still in beta.
Can you give me an example of how you would use the data?
GR: There are many specialized parts of our business so one of the things we’ve started doing is looking at where in our business this kind of data can add real value. Real value comes when you can start to match different data sets together. Let’s say you take a client like Budweiser and they have a campaign in the Super Bowl. It would be interesting to see through carrier data what kind of impact any Budweiser advertising has had when people are browsing their phone or going on Facebook. We would then combine that with other data that we might collect to come up with a total media exposure.
Another example could be that when we’re collecting consumer insight from our surveys, we’ll ask people questions about the apps they remember seeing, what products they’ve been buying. It would be interesting to match the mobile carrier data with our survey data to see whether the effect we’ve measured in the carrier data is reflected in the surveys. Do they remember having seen those ads?
How are you addressing the privacy concerns around companies monetizing mobile user data?
GR: It is a sensitive area but a lot of the outcry has not been based on all the facts. What SAP is talking about is anonymized, aggregated data. We’re only interested in what are the top 100 apps according to hours spent on your phone or the most viewed adverts on your phone. That’s the kind of thing we’re interested in.