Exelate’s Zagorski: Walled Gardens Go Against the Interest of Marketers

Mark ZagorskiExelate, a data technology company, announced yesterday that their existing product suite – which features a DMP, a data marketplace and a multi-faceted analytics solution – is to be unified in a single cloud. This is an interesting time to watch the data technology companies:  on the one hand there are fascinating new frontiers opening up for data-driven advertising, such as audio and TV. But on the other hand many of the industries leading players are building walled gardens, where they control the data, the inventory and the technology.

Mark Zagorski, CEO at Exelate, says his company is “railing against” the walled gardens. “We think that those environments aren’t good for marketers for many reasons,” he told VAN. “They’re limiting the marketers, so it’s like a product manufacturer being forced to sell all of their products through Walmart, which when you think about it is a walled garden in the retail world. While Walmart are big and have lots of access, they don’t have access to everybody in every place. So if you put your product in a walled garden, it may seem very comfortable at first, but ultimately you are trapped. You get tied in, you lose your ability negotiate on price and terms. Your data is owned within that garden and you won’t be able to take your data outside of that garden and apply those learnings to other media.”

Zagorski feels these trends go against the spirit of openness that was more prevalent during the early days of programmatic and RTB. “Exelate plays in a world which is very open and we’re plugged into 125 different platforms, and we’re pulling data from hundreds of different sources. This flexibility and openness provides huge advantages to marketers to find out what does and doesn’t work. The reason we have a DMP as part of our cloud solution is that it’s vital for marketers to be able to bring first party data into that cloud, to be able to mix it with third party data, to activate it across any platform they want to, and then measure it in an objective way. Remember we don’t sell media and so we can afford to be agnostic. When you look at walled garden companies like AOL, Facebook and Google, those companies are trying to sell technology, measurement and media as one package. So how are you ever going to get a straight answer from those companies when it comes to the effectiveness of media or the technology when you’re not allowed to move it out of their space. The flowers are always prettiest in a garden when you can’t see flowers anywhere else.”

Emerging Opportunities in TV and Audio 

However, in spite of Zagorski’s  misgivings about wider industry trends, he remains bullish about some of the new fronts Exelate are working on when it comes to creating unified user profiles, particularly when it comes to audio stream targeting and TV data. On audio streaming, Exelate have built out the first audio DMP in partnership with Triton Digital, which focuses mainly on audio streaming via tablets and mobile devices.

On the connected TV side, Exelate have been partnering with Innovid and Roku on connected TV targeting. In the past, smart TV targeting has always encountered problems. While cookies work on most smart TVs, the reality is that very few users use their TVs to browse the web, so intent-based data sourced directly from connected TVs is almost non-existent. However, Exelate are now able to connect activity that comes from a mobile device directly to a Smart TV device. So if someone is looking at an automotive site on their mobile device, that data can be immediately forwarded to Innovid in real-time, allowing them to target the Smart TV advertising according to what is being consumed on the mobile.

This cross-platform capability enabled by unified customer profiles, which Zagorski says depends on ‘billions of signals’ ingested by Exelate from hundreds of sources each day. “All of the signals are coming from mobile devices, from online interactions and from offline interactions. So we’re importing all of this data into our cloud. The first thing we look at is deterministic, so for example we might look to see if we have onboarded the same user via some sort of hashed PII across multiple devices. So we may have the seen the same user for online activity as well as mobile activity, which enables us to synch those up,” he said.

“The second are algorithmic decisions, which combines the best of device matching with algorithmic predictive matching. So it looks at whether we’ve seen the same consumer accessing data from different devices from the same wifi address. So, for example, we might look at someone using a smart TV application and map the wifi address of the smart TV box to the mobile device, which tells us that there’s a very high chance that the person watching a smart TV video is the same person that’s browsing content on their mobile or tablet device. So if we look at the probability of these two devices being the same person based on other activities we’ve seen, we can make a very highly educated guess that this is the same person we’re seeing,” he added.

Zagorski says that this type of approach is superior to the road taken by the addressable TV companies. “Those companies have been around for years and they have attempted to do some super-heavy lifting i.e. to get into the linear TV stream, although it has taken them decades to make progress. I don’t think that focusing on linear and changing the ‘heavy iron’ is the way we’re going to change TV,” he said.

“Now that people are consuming video in a very different way and video on demand is having such a huge impact, the change is going to come from things like Smart TV applications and then moving into the set-top box and focusing on time-shifted television. This is far easier than changing linear TV by pulling in data from the set-top box by pulling in data from the set-top box, sending it back to the cable company and changing the ad that is being served at 9.03pm during a programme, all of which I think would be a huge challenge and a waste of time,” he added.

Subscribe to Weekly VAN Newsletter

Ad TechConnected TVDataProgrammaticTV