AdExchanger: Do you work with agencies as well as brands?
ERIC STEIN: We just started an agency team and an agency practice in the past [four] months or so. That’s on the media side of the business. We’re working more with agencies on the data strategy, the services side of our business, helping them incorporate data into their plans for mutual clients.
What does that work usually entail?
We help them understand the difference our data can make. As agencies get more savvy, we see the opportunity to help agencies understand the data that’s available and how it’s different from one provider to another.
When you think of an Acxiom or Experian, our data compares favorably. We did an audit not that long ago that showed our data can drive better performance than any of the data sets available out there from [four] different competitors, Acxiom and Experian included. [Ed: The other two were KBM Group and Infogroup.]
When you incorporate data, is it just Epsilon data or will it include data from competitors as well?
We can help them with whatever data they have at their disposal.
Who performed the audit?
I don’t know if we’re allowed to mention the name of the company because they have relationships with all the other people who were audited.
Was it an audit Epsilon commissioned?
We commissioned it but it was independent and industrywide.
What are your clients most concerned about?
It varies client by client. Many start out needing a plan to show how they get the different silos of data together in one place, so they can have the broadest impact on their marketing efforts over all. And strategizing some of the tactics that will allow them to execute those programs effectively. Those are some of the primary things we start out with.
Then it’s helping them decide what technology to use and/or getting the most out of the tech they already licensed. The big differentiator for us is we haven’t gone and built proprietary technology, like Acxiom announced Audience Operating System. We partner to build most of that and we think leveraging best-of-breed technology to help a client is a more viable position for us in the market than building a technology platform to compete with the Adobes and Googles.
Who are your partners?
We’re one of the leading members of [MediaMath’s] certification program. We run the campaigns that we manage through MediaMath. When it comes to data storage and DMPs, depending [on client objectives], we work with a variety of online platforms, some of which you’d call a DMP, like a BlueKai or a Turn or MediaMath as well. And others that are not necessary classified as DMPs but are online platforms like an Adometry.
Do these technologies perform notably differently than their competitors?
There are minor differences across the board. Some of it is a pre-existing footprint with a client. Adobe has a pretty strong leg to stand on in that regard. Some of it, like an Adometry, is the specialization in the analytics space. That’s an important issue as clients aggregate data from different silos and try to perform the analytics as the customer journey has become less linear, more fragmented. Trying to pull the data back into the platform and understanding the impact of your marketing activities is a critical, critical need.
Can you go into how different clients want their data sliced, and how that influences the tech partners you work with?
If you have a very media-oriented client that wants to execute campaigns across different exchanges, MediaMath – and that’s why we chose them – has the best reach available. You hear clients talk about leveraging three different DSPs or what have you. We can help [our clients] think that through and we believe that incremental reach you might get is negligible versus the complications of executing those campaigns to different audiences across different platforms.
When you’re in a more analytic-oriented role, someone like Adometry can make pretty good sense as a platform. We run some analytics through MediaMath’s platform as well and they have good features and functionality for campaigns run on that platform.
And the site personalization opportunity is Adobe’s bailiwick, and they’re differentiated with their footprint and getting access to that data from the Omniture acquisition. Those are some of the major differentiators among those players.
Turn and BlueKai are also well-established. We’ve used those guys with some clients who have already licensed their technology.
How do you link online and offline? What are the complexities there and what can and can’t be accomplished?
There are some fundamental pieces: Scale is one of them, and so are access points.
We’ve invested a lot of time and effort in building our own reference file, which is what we call it internally. We can take a client’s data and add attributes to it to drive higher match rates. We have lots of data attributes that enable us to enhance a client’s file to drive those metrics.
On what’s achievable, it depends on client objectives. If your objectives are to reach your audience across the various silos of the Internet in which they might be reading content or engaging, then that’s a critical focus in terms of building partnerships and capabilities to distribute data out to those silos.
The integrations with Facebook and Twitter are proprietary integrations that leverage more of those attributes we have in our reference files to drive match rates.
Facebook has Custom Audiences and what we work on with Facebook is called Managed Custom Audiences, and we use 10-plus different attributes to drive match rates. Whereas the publicly available Customer Audiences tool only uses email and phone number.
How are match rates evaluated?
A match rate is the product of a numerator and denominator. You’ll find people talking about very high match rates, in the 80s or 90s percentage area. But that depends on where you start. If I start with a full client CRM database where they have less than 100% coverage of email, maybe the Fortune 100 and beyond have somewhere between 40-50% of email coverage in their CRM or loyalty programs. If you start with something like that, we can drive higher match rates that push us upwards of mid-40 percentages on average, so that would be a pretty good match rate.
When you think of scale of user IDs, that’s Facebook and Twitter. Those platforms would be who you’d focus on.