With a PhD. in econometrics and an analytics and statistical background on the AdAdvisor side of the Neustar “house,” VP of Research & Analytics Ken Inman has good reason to be laser-focused on building consumer predictive models. Inman oversees the development of targeted marketing applications, using data and information to help address challenges in customer relationship management (CRM). In other words, he’s trying to understand what motivates people to buy.
Inman says, “Predictive modeling and trying to figure out a target audience for whatever service or offering you provide – that’s what my world is about.”
Last week, Neustar announced a new analytics tool called “Campaign Conversion Insights (CCI),” which the company describes as a “service that allows marketers to measure the performance of digital ad campaigns based on conversion, leads and sales.” Read it.
AdExchanger spoke to Inman last week about the new product, Neustar AdAvidsor and industry trends.
AdExchanger: Given your 16 years of experience, how would you say digital has changed the way predictive models are created or built?
KEN INMAN: Initially it was all about using data that you could see in an online channel to drive predictive models. But now it is about trying to fuse all sorts of information together to better support online audiences.
One of the biggest things that people care about today is measuring outcomes or “How effective are my online campaigns?” And not just measuring by click-throughs or things you can easily observe but measuring according to what the marketer cares about – somebody buying as a result of a display campaign, for example. And then you have to do that in a way that’s privacy-sensitive or privacy-by-design.
Sixteen years ago obviously we didn’t have all this. I remember we had something called a “Gopher,” which we thought was an easy way to navigate data sites, and it was between academic institutions.
So in trying to understanding "outcomes," is it best to use a top-down approach or a bottom-up approach? A bottom-up approach – at least from my perspective – is starting at the outcome and working backward as opposed to beginning with the tactics.
I’ll say we bring it all together in the middle, which leads to our Campaign Conversion Insights (CCI) offering that we’re releasing. At the bottom, there is authoritative customer transaction information. At the top, you’ve got your online media campaign running through some display advertising campaign platform. What you want to do is bring those two together in the middle so you can look at whom you’ve shown ads to. And you want to combine these two in the middle so you can truly say, “Here are the millions of folks that fell into the audience and here are the thousands of folks that were in that audience who then followed through with the purchase.”
So we’ve developed partnerships with media platforms for display campaigns. We’ve also developed relationships with authoritative consumer transactions data [providers]. We’ve got the ability to collect that authoritative transactions data from offline sources and other areas, and we’ve developed techniques to bring those together in a privacy-by-design fashion that doesn’t infringe on your personal identity or profile you specifically, but is more about understanding you as part of a group. If you look at how audiences or how campaigns are verified and validated today, it’s not very often that they’re verified and validated based on the actual outcome, which is what you really care about. We’re providing a link.
What does authoritative transaction data mean?
Working with companies that actually have access to purchasing behavior.
Is this like a traditional “lead” scoring methodology that’s now a part of CCI? A new take on scoring, perhaps?
With scoring, there are some similar elements, but it’s different. Scoring is bringing data and information to bear so that I can put you into an audience. CCI is looking within an audience and tying you to some kind of other external group. In a sense, you’re fusing data together to solve a problem. To that extent, yes, they’re the same. In this case, the objective isn’t to score you, though; it’s to combine some data sets – an information set about purchasing with an information set about ad exposure.
Can Neustar bring some of its own data sets to bear beyond what the AdAdvisor unit has traditionally focused on?
In this product, no. Neustar has to be very careful, because a lot of the data that they have is protected. It’s a contributory database that cannot be used for anything other than a specific purpose, which is allowing people to move their phone number from one provider to another and determining who owns the call and how to route a phone call. If you’re talking about one of the core Neustar data assets, no.
On the other hand, there are some other Neustar data sets like Quova that help us do some of our work. They maintain information about IP addresses and have built a commercial product where they’re licensing information associated with an IP address – who can own that IP address and other attributes, is it on a mobile gateway, things like that.
First-party data versus third-party data: how that might work with CCI?
One of the things we’re doing is empowering companies to bring their first-party databases online so they can be used for audience creation. There are obvious restrictions too. But it’s definitely something that we’re facilitating not just for CCI but as an audience creation tool. If you bring your first-party data online, then you can have a consistent messaging to your customers, whether you see them in an online environment or in a call center, because you can also use our tools and APIs to target the same audience. Now we have analytics infrastructure and an audience creation process that allows you to find that audience regardless of channel.
Lead generation world versus the audience-buying programmatic media world: what are the key differences in terms of predictive modeling needs? Or where is the overlap?
They definitely overlap. When I think about the lead gen business, you’ve got somebody who needs to reach the same audience over and over and over, and it’s a very specific definition; it’s like “Build me a specific, custom audience.” In the lead gen space, we work with unique companies that are spending a lot of money in purchasing leads, and they need scoring models very specific to what their business is all about – so, they need very customized audiences.
In the programmatic space, we’re working with companies to provide an audience quickly, because every week companies change their targets, or every week they’re running a different campaign. It’s a very dynamic environment they’re operating in. You can’t spend in the same way, trying to build that one very customized solution that runs and runs and runs as it does in lead gen. Every week there are new campaigns in programmatic. That’s AdAdvisor’s space.
Does CCI address the issues in a mobile, specifically around identification?
Mobile is a little different. We definitely have a lot of what it’s going to take to be successful in the mobile space, but there are some unique challenges in mobile as well, just because of the nature of the devices and your ability to identify them in a consistent way. We have the supporting infrastructure to provide insight, but the ability to identify these mobile devices in a privacy-sensitive manner is not quite there the way you can fuel it in a more traditional online space. It’s coming, though. It’s got to get there.
How about CCI and Connected TV?
Connected TVs and addressable advertising is on the roadmap. We have a lot of relationships with all the major MSOs. Today, not many people have those kinds of devices and capabilities. You’ll surely be able to do the same things with that data in terms of tying it to that offline authoritative transactions sources, and you can view what we’re talking about here with CCI. It’s probably a couple years until there’s sufficient scale in terms of the connected TVs and the addressable advertising space out there.