Tnooz had been slated to do an interview with Cambridge Analytica’s Chief Data Officer Dr. Alexander Tayler during this year’s tnoozLIVE@ITB program. Yet the scheduled time slot came and went without the appearance of Dr. Tayler. Given the firm’s footprint in travel, with clients ranging from a large global airline and major hotel chains, this interview promised to offer insights into the firm’s work in the industry.

The firm has rocketed to prominence, making it quite the interesting interview to cover. As Dr. Tayler himself said from the stage at ITB Berlin this year, the firm is now famous: “If you get Donald Trump elected, all of a sudden you are a household name.”

As we weren’t able to explore any questions related to personalized advertising in travel and given the recent allegations that the company used un-permissioned user data from Facebook in its efforts to personalize content and communications strategy, it felt right to review Dr. Tayler’s on-stage Executive Interview through the lens of this evolving global narrative around data, privacy, and psychographic manipulation on social media.

Quotes have been edited for brevity and clarity. The full video is at the end of this article, with clips set for each subheading.

From programmatic to 1:1, from segments to individuals

The firm’s approach is what really propelled it to prominence. It’s a complex process that moves beyond programmatic to deploying deep profiling to build personalized creative for consumers.

Here’s how the firm operates, offering some insights into how data can be used to personalize content and communications across channels. It’s a long answer, but it’s quite fascinating to understand just how complex the machinations are to pair a targeted segment of individuals with creative.

You’re still addressing segments, but each segment is made of individuals. Then it’s about putting individual people in the right segments with people that are similar to them. So when you do go off a campaign, you’re giving people creative that they’re going to engage with.

At a higher level, it all starts with data and research, its first party primary research such as designing a survey and recruiting a sample or a client’s email list, match those data back to other datasets, what we call very deep-but-narrow data in their own CRM, in their own inbound/outbound marketing platforms. This tends to be data that cover every touchpoint that client has with their customers.

When those very deep and narrow first-party data are put together with what we call very broad-but-shallow demographic and socio-economic and consumer lifestyle data, which can come from places as disparate as licensing it form Axiom or Experian (who get data through their own network of partnerships), or even things as simple as enriching your data at some geographic level of granularity with Census information. That brings a lot of breadth and context about individuals.

With the combination of deep-and-narrow and broad-and-shallow, the contextual, with the research that gives you something actionable and insightful. Then you layer that on top, so from the combination of three different independent sources, you’ve got all the ingredients you need to use a machine learning algorithm to do something like extrapolate out some of the insights from the research to know which segment you ought to place people in.

Having done that extrapolation and segmentation exercise to create audience groups, you have a team of creatives interpret what the most effective way to communicate with each segment is. You’re starting with a different set of ingredients that gives people a completely different experience in engaging with your brand.

If data is the spear’s shaft, ‘probing’ is its sharp tip

The secret sauce that has powered Cambridge Analytica’s success is also what caused the recent discomfort from how the firm used dubiously-collected third-party data — behavioral science. The company puts psychologists alongside creatives to better target advertising for clients.

While this isn’t really a new thing in advertising, the psychographic profiling made possible by today’s data-rich world is unlike anything ever before. Thanks to low data storage costs and sophisticated data science techniques, micro-targeting has closed even tighter. It’s now all about 1:1 marketing, or a brand speaking directly to an individual consumer. Or, as Dr. Thayler puts it, the firm “probes” the consumer psyche to pair creative to consumer.

The big differentiator we have is the behavioral science. We started life as a behavioral science thinktank, and these academics were looking for ways to commercialize what they were doing in applied psychology. So we’ve still got this very strong basis of intellectual property in the company about the behavioral drivers that make people respond to the world around them in the way that they do.

We’ve got a team of psychologists that sit within our research division. When we’re putting together a search instrument to probe which people are likely to engage with what sorts of messaging or product, they will generate a hypothesis about what psychological traits are going to be very powerful or very useful in this particular study.

We will probe those alongside the more conventional behavioral or attitudinal questions to derive quite a very fundamental basis for audience segmentation.

This also gives us very powerful clues in terms of knowing how to cut the creative for those particular audiences.

But, as Dr. Tayler rightly emphasizes, a bad product will never win. It might get some initial traction, but without a solid product, no amount of targeted advertising will work.

It’s not about tricking people into buying something they otherwise wouldn’t.

The truth is you might be able to run a [successful] trial campaign, but if the product is no good it won’t stick. So you might be able to driver shorter term engagement by doing something clever but the money would be better spent fixing the fundamental problem.

Data should be owned by the user as a monetizable asset

Dr. Tayler believes that user data should actually be owned by the user, with the ability to leverage into exiting commercially-syndicated sources of user data:

Already products are popping up that will allow people to syndicate their own data and share some of the wealth that has been generated from data-driven targeting.

The opportunity to get more data up and out of commercial data silos and into people’s hands is an opportunity to help people derive insights into their own data. And to have avenues to monetize those data.

The company doesn’t own any of its own data

Given the ongoing turmoil around his company’s alleged misuse of ill-gotten Facebook data, Dr. Thayler seems to suggest that the company actually owns no data. If the allegations made public this week are true, there’s some clear confusion on what happens when a firm uses data it has purchased to integrate into a core product that may expand far beyond what the original data owner had contracted.

We don’t own any data of our own, we’re a consultancy. We work with data from different sources.

As mentioned yesterday, it’s up to all travel brands to understand and vet their partners so that there are no questions about the legitimacy of data being used to target travelers.

e-Commerce: Split testing is out, behavioral is in

Dr. Tayler offered strategic guidance to any brand looking to perform better on the e-commerce front. It’s rare to have a vendor explain in such detail what it does for clients. So there are some insights here, as far as how Cambridge Analytica outperforms for its clients. It’s also littered with minefields for the novice data-driven marketer.

There’s been a lot of research over the past couple of years that show that people don’t think companies recognize them as individuals anymore. And a big part of that comes down to the fact that these companies are literally treating them as being numbers in an optimization routine.

We’ve got a team of behavioral scientists that contribute to the research design as well as the interpretation of the outset of the output. This helps our clients understand what image and what tone and what narrative to use to engage these different segments.

So what we do as a company is we go further up the funnel. We start with a program of audience research and combining data at the respondent level with data from different sources (client data, other publicly or commercially available third-party datasets), and then we use this to get information about different audience groups. What they want and what they need, what messaging they’re going to engage with and then we structure a communications campaign on that basis.

In that way, we’re avoiding the habit of randomly testing creative against different audience segments to see what works. Because you’re taking the time to understand those people before you try to communicate with them. And that’s really what our Modus Operandi is as a company.

The opportunity for more data-driven attribution in travel

This is an interesting point, as we’ve asked the question about how attribution plays out in the future of travel media. As noted, attribution offers deep insights to travel marketers as far as which segments respond to which offers on which channels. Again, caveat emptor to marketers: with GDPR and general increased awareness from users, be thoughtful and smart about how you deploy user data!

There are a few reasons why [advertising] is opaque, which creative and segments are converting and which ones aren’t. It’s down to the fact that you’ve got a whole myriad of different platforms and systems between the advertiser and the audience. Being able to get accurate attribution data, or knowledge about what’s working and what isn’t, is what’s required.

What’s really needed in an industry like travel, is that those businesses who have got a data offering this cross-market perspective — things like price comparison or booking websites — to syndicate those data about which clicks are turning into conversions back to the advertiser. So then they know what is working and can optimize their campaign on that basis.

More and more, people who are in possession of very rich and very broad attribution data, are starting to be aware of the value of those data to a marketer. And the opportunity to monetize those insights is a very compelling thing. It’s unlocking data from these different silos.

I’m sure it’s going to happen in travel sooner or later. Hopefully sooner.

Implications for GDPR

While it might seem that GDPR’s right to data portability, as well as its stringent data control requirements, would threaten a business that relies on the free-flow of data from commercial partners. Not really, as there seems to be more optimism around the ways that users will now be able to engage and share data with companies.

GDPR is an enormous opportunity for a company like ours.

Regulatory compliance isn’t the sexiest topic for most people to think about, but it can’t be understated what an enormous change this law is. The biggest implications are things like requirements around rights to access, data portability, and right to erasure.

This means that any company holding personal information, or information that an individual can be identified by (not necessarily name and address), if a user can be uniquely identified, that’s classed as personal information. Users will have the right to say to request a copy of that data in a machine-readable format, and be removed from the database. So the data now becomes yours to control and manage and monetize.

This means that a lot of companies making money from monetizing other people’s data are now looking at a fundamental change in the way their businesses operate. Everyone from supermarket loyalty cards to the world of ad targeting.

Under GDPR, cookie IDs and mobile device IDs are being classified as personal information. And you will have the right to go and reclaim those data and stop them from using them to target you.

As Facebook is mired in a scandal around allowing out-of-policy data scraping on its own sites, there could indeed be a windfall as users take control of their data to deploy in independent data syndicates.

The ongoing conversation of this Facebook/Cambridge Analytica issue will not only increase awareness of data control among users but also embolden companies that aim to offer a product for the post-GDPR age. It’s a brave new world, and data control, data privacy, and data security are firmly at its core.

Watch the video in its entirety on the ITB YouTube channel.