An Interview with Ken Dreifach
At SpotRight we deal with consumer data every day and put great emphasis on its ethical collection and use. This is one reason why we are delighted to have Ken Dreifach as an advisor. Ken is one of the foremost digital data privacy experts in our space, with significant regulatory and business experience. We recently spoke to him about the increasing focus on people-based marketing.
SpotRight: Ken, you have such deep experience with online data and internet privacy. How did you get started on this path that made you an expert in the law related to emerging technologies and online privacy?
Ken Dreifach: Back in 2000 I was working for New York State Attorney General, leading The Internet Bureau. At that time the first online data “scandal” started to unfold, when DoubleClick (pre-Google), which was essentially an ad network collecting data and serving ads on thousands of sites even back then, announced that it was going to buy Abacus, which was and is an offline data company. They indicated that they planned to merge offline data and online data, and this got privacy advocates and regulators very much up in arms because people were concerned that anonymity online would effectively be lost. In other words, that personal information would be merged with what we call non-PII (personally identifiable information) or web browsing data.
This was really the first case that set up these rules that said online data and offline data — except in certain circumstances — shouldn’t be merged. In other words, the vast pools of anonymous third-party browsing data shouldn’t be personally identified from an anonymous state.
That was my introduction into this whole space. It’s interesting to me that ten, 15, almost 20 years later, we’re continue to deal with some of the same issues – figuring out how, when, why and how it’s okay to let the online and offline world, the offline identity of a person and the online identity of a person, converge.
SpotRight: It’s surprising to hear that that was your entry in this space because these are still questions we’re trying to resolve now. Where is that line of what is ok and what is not?
Ken: Exactly. There are lots of signs that people — consumers, app users, social media users — view their privacy differently ad they have a more open sense and a more accepting sense of the value of data and targeting and personalization of ads and content. But every so often we still have to take a step back and ask where we are and rewrite the privacy rules of the road a little bit.
SpotRight: Speaking of rewriting rules, we’ve seen a shift from data-driven marketing to people-based marketing. Do you see the same thing, and what’s the difference between the two?
Ken: I would say person-driven or people-driven marketing is a subset of data-driven marketing. But there’s definitely a very significant trend in the last five years or so, which has sped up in the last two years, to use offline data – what we have always thought of as offline personal information – in the online ad and content targeting and analytics space.
A lot of what’s driven that is the emergence of platforms like LiveRamp and Neustar which are able to build cookie pools that are derived from personal information and to use those identifiers to sync data across lots of different platforms.
Those kinds of technologies allow personal information, whether it’s social or email or other channels that are traditionally thought of as involving personal information, to either be merged with online delivery systems or to be overlaid against online data, in ways that respect privacy.
So, when we talk about people-based marketing, we’re often talking about marketing that has some component of that targeting or distribution system, where you’ve got online targeting that was derived from a real person. It’s a more precise type of marketing.
When we talk about people-based marketing, we’re often talking about marketing that has some component of that targeting or distribution system, where you’ve got online targeting that was derived from a real person. It’s a more precise type of marketing.
SpotRight: To that point, when we talk about person-based marketing and executing on it via online platforms, what is a person to a machine?
Ken: Well, I think to a machine, a “person” is a unique identifier that is tied to a truly unique individual identity. In other words, it’s not inference that you might be somebody who might be interested in buying movie tickets on a Thursday night. Rather it might be tied to an actual individual or a persona that we know does buy movie tickets on Thursday night. So, in the ad tech context, it’s more what we call deterministic and less purely inference-based.
SpotRight: What are the privacy implications of this, knowing that it is deterministic and tied to a real person?
Ken: The privacy principle is really the same point that we started off with – with the DoubleClick Abacus example – which is that we want to make sure that even though we’re deriving online data from personal information, we’re not taking information that reasonable people think is anonymous, which could be browsing information — it could be location data, where you happen to be with your phone at a given point — and converting that into genuine readable personal information, in a way that might surprise a consumer. That’s the main privacy concern.
There are some other privacy points. We want to treat information regarding a person’s health condition with sensitivity. We want to make sure that consumers have an accessible place to go to learn how their information is used, privacy policies, and ways to opt out. Most of the big platforms honor these privacy principles, and industry associations invest in widely available solutions.
SpotRight: That makes sense. What benefits do you see consumers reaping from person-based marketing?
Ken: All of this leads to better marketing. It leads to consumers getting more relevant ads. It gives marketers ways to measure whether or not consumer like their ads, because you can tie your ad campaigns to actual real transactional data.
We always say that better targeting leads to better ads, which leads to happier consumers. Consumers are less likely to get ads they find annoying; they’re more likely to get ads for products that they like. And, at the end of the day, they probably get fewer ads because the ads are better, and the ads are more efficient. There’s more leeway for all of the companies that are relying on advertising to pay for their media. There’s more opportunity for them to improve and show fewer ads.
SpotRight: How do you see person-based marketing evolving in the near or medium-term horizon?
Ken: I think we’re going to see more person-based customization, and I think it’s going to go beyond ads. I think it’s going to go to really all types of commercial practices. I think person-based analytics is going to be used increasingly in product design. And I think we’re going to see person-based customization built into products.
We’re seeing connected cars, connected appliances. We are starting to see connected cities and neighborhoods. In New York some new very large-scale urban developments are going up. And many aspects of the urban environment are wired, commercial, for both retail, and homes.
The ability to create these identified identities and personas and connect them across platforms, home, work, neighborhood, online, mobile, social, etc., has potential that I think we’re just starting to tap into. It’s starting with advertising, but the potential of a person-based analytics goes beyond that.
SpotRight: Do you think that some of the challenges that advertising has had to overcome, to think about and determine what is an appropriate use of data and what is not, that will be used to help inform some of these future developments, like connected neighborhoods, for example?
Ken: I think it will. One thing that the advertising and marketing world has been very good about is creating user notice and choice. Five and ten years ago there was some criticism heard on this, but now opt-out messages and succinct explanations of how data is being selected and used are ubiquitous.