AdExchanger: What do your clients want in terms of cross-device customer insights, and what can you provide?
JOHN NARDONE: We can give complete cross-device connectivity when it’s for customers of the clients. If you take one of our banking clients, we've linked the device IDs across multiple computers and devices for their customers, because you can complete the linkage through a login. You log in to online banking tools through your computer at work and at home. You answer an email on your mobile device. We’ve linked that together and that links back to your account number.
Once you’re in the realm of CRM, this is broadly accomplished for our clients, and the DMP facilitates that. You’re just managing the database of the various device IDs and cookies associated with the master client Key. You go into our DMP, there are multiple options with how you want to key your data.
What do you mean by “key your data?”
What’s the master ID around which all the other pieces of data that are connected? In a standard, old-school DMP implementation, the cookie is the key to which all the other data is attached. When you go into our platform, you can look at it through lens of the [x+1] cookie or through the lens of the client key, an encryption of the customer’s account ID.
What we’re launching in April is the offline key, a key to the national household database that links to database marketing services providers. So there’s a persistent key linked to you as a human being and the household you belong to vs. your account number at the bank or your cookie.
The key is the connective tissue linking customer activity across devices and channels?
The key is the connective tissue to which all the data is connected. Our DMP has a unique architecture in that it allows concurrent keys at the same time. You use different keys with different situations with different partners. If you do prospect acquisition in display advertising, the cookie becomes the relevant key. If you do targeting for owned touchpoints like CRM, then the customer ID becomes key. If you do targeting for offline channels, like building segments for direct mail or email – which bizarrely is online but it behaves like an offline channel – then you’re using the persistent offline key as the relevant key. And they’re all cross-linked.
How do you make that ultimate identification?
We’re blessed with a lot of client connections where customers identify themselves.
What about when you’re dealing with anonymous consumers?
If it’s anonymous, you get partial linkage. Maybe you can link offline with a cookie, but you don’t have a customer ID. Or you don’t know if they’re a customer because they’re anonymous. Or you might have a mobile ID linked to a cookie, so we know it’s the same person. Typically that linkage happens over email, so maybe someone accesses email over the mobile device and desktop. Then you can link mobile ID with a desktop cookie. We know it’s the same person but we don’t know who the person is. You end up with a web of partial linkages that we piece together as a puzzle.
How accurate is this? What’s a good accuracy rate?
I don’t know if you can answer that. Good accuracy rate for what? Good accuracy rate for what purpose, what channel, what client? It’s all over the board. It depends on what sort of asset that client has tried to build. If you build your linkage as an asset as a client, you have to make a decision to do so. This means tagging your emails, which is a very simple because it builds your match footprint. Regardless of what it is at any given moment, it’s increasing all the time.
Where doesn’t accuracy have to be high?
Branding campaign, accuracy is not that important. But if you’re giving credit offers, accuracy is absolutely essential or else you’re breaking the law.
In terms of technology, what do you need to do on behalf of creating that linkage, especially with filling in the anonymity piece?
It’s not a question of technology. That’s not in question. There are things you won’t do because it crosses the privacy line. For example, we will go from known person, known ID or offline ID to online targeting, but we won’t let it go back the other way. We won’t let people say, “I have a cookie, tell me who it is.”
There’s nothing technologically keeping you from doing that, but it’s bad policy from a privacy standpoint. Just because the technology exists doesn’t mean we should do that.
Is that legal or self-regulation?
There’s no legal anything in this place. There are self-regulatory guidelines, which shouldn’t be the definers of common sense. I think that self-regulatory guidelines are just that. They shouldn’t be something for people to hide behind. You have to do what’s right and what’s right is a question of what’s the nature of the relationship you have with this person.