Data management is like dental hygiene: No one enjoys doing it or even thinking about it, but you’ve got to stay on top of it or else the situation will get really messy really quickly.
Joanna O’Connell, CMO of MediaMath, on Thursday moderated a panel about data management that felt like a good cleanse.
Here are the highlights, as spoken by Darryl Gray (SAP Exchange Media VP of sales and business development), Anneka Gupta (LiveRamp/Acxiom chief product officer), John Gallagher (IBM VP of performance and programmatic marketing), Seth Demsey (AOL Platform CTO) and Ali Bohra (Adobe Audience Manager/Advertising Solutions director of product marketing).
Dealing With Data Fragmentation
GUPTA (LiveRamp/Acxiom): I’m consuming content across a multitude of devices, interacting in the real world, and all of that data being collected is tied to a different kind of identifier. To get a complete picture, you need an identity layer, which needs to understand who I am in any way I can be identified: by device, email address, phone number, through Xbox or Roku. It’s a very challenging problem. …
There’s so much data, you have to figure out what to do with it and gain the insights. You need a hypothesis about what you’re testing and the result you want to get. The data can help prove what the hypothesis is – but the data won’t show you an answer without you applying what you know about your customer.
Reconciling Privacy
GUPTA (LiveRamp/Acxiom): While identification can help provide meaningful experiences, it comes with a sense of privacy challenges. What usage restrictions will you place in terms of how you architect the data solution and how data flows through all of it?
GALLAGHER (IBM): You have to know want you want to do from a data point of view. Historically, businesses are about capturing transactions, not individual data. So you need to reset your logic on data. This is not an enterprise transaction; this is an individual interaction.
And how do you use that from an analytics perspective … and always with the context of transparency and purpose.
[The industry is] not there yet. It’s very clear we’re not.
The Publisher Problem With Multiple Identities
DEMSEY (AOL Platforms): IBM will have their own view of the client, AOL will have their own view, LiveRamp will have their own view. Media measurement companies all have their own views.
Publishers are challenged with inconsistencies when they’re being verified and audited [by companies that] don’t know beforehand the difference between the targeting, planning and measurement. When publishers serve an ad they could’ve served to somebody else, only to be told after the fact, “You know what? That wasn’t a 19- to 34-year-old male, so we won’t pay you for that,” that pushes price pressure back onto the publishers, who then raise prices.
We’ve all benefitted from loosely coupling our technologies together to create powerful chains of capabilities. But we’ve failed at figuring out how these things work together, and how they don’t line up.
And when it takes nine calls to nine vendors and seven seconds to start an ad – that’s why people put ad blockers on.
Is the client browser the appropriate place for these things to be done anymore? I don’t know.
Using Combined Data Sets
BOHRA (Adobe): Third-party data used to be key to understanding the consumer. Second-party data – a private exchange of data between two partners – has become more prevalent in the space. From an Adobe perspective, we have marketing and publishing customers trying to figure out how to leverage it.
Not every marketer or publisher is trying to monetize their data. It’s specific to certain verticals. In the TV space, there’s a really interesting ecosystem between the MVPDs and operators like Comcast and DISH, and the broadcasters, which have a lot of data around linear and addressable viewing. And there’s the marketers trying to buy spots in that space. That ecosystem lends itself well to sharing data between partners.
In 2017, you’ll start to see material ways for marketers and publishers to access streaming and viewability data from these large operators and MVPDs.
Another vertical is travel and hospitality. There are a lot of synergies between hotel chains, airlines and even credit card companies. These large airline and hotel brands are sending customer records and the actual device and cookie IDs, which let them understand the behavior between one entity and another.
GRAY (SAP): We’re still not leveraging first-party data the way we should. We’ve had tons of clients who’ve told us that they’re not unlocking the value of the first-party data in their system.
Later in the year, we’ll release a DMP focused on first-party data, leveraging all information in your customer engagement systems and building out your POS and ecommerce data. Then it brings collated third-party data together with first-party data in real time and extends that to a trusted second-party environment.
With second-party data, you need partners willing to work together and you need a trusted environment, and that’s something we’re building out. Under Armour collects a lot of fitness and dietary information. Certain companies, like those that sell beauty products, would love to have that information.
Using Verizon’s Data For Marketing
DEMSEY (AOL Platforms): One thing that was so exciting about being part of Verizon is the opportunity to figure out what to do with a scaled, proprietary data asset. Now, everyone has a scaled, proprietary data asset. The exploration we’re going through is how do we advantage end users, publishers and advertisers? How do we use this data to create better experiences? How do we become more effective with our advertising campaigns?
We can all do that with our data. It takes giving yourself permission to look at the data assets you have and stitching them together in unique ways. Verizon has network data, television data, CRM data, coupled with ad exchange, DSP, ad network, mobile app and publishing data.
We came up with some interesting concepts – not just device graphing. We wondered how it impacts frequency capping or bid prices or message sequencing. How does this effect attribution in home viewing?
Then we can stitch together this end-to-end user journey by combining our data footprint and the footprint our clients bring to us.
