Dentsu’s Chief Data And Technology Officer On Generative AI, Cookies And The Future Of Data

AI often steals the spotlight these days, but it is nothing without a data set.

In a nod to the interconnectedness of data, the agency holding company Dentsu created a new data and technology group. And in March, Shirli Zelcer stepped in to lead the group as Dentsu’s first chief data and technology officer.

In her new role, Zelcer oversees three areas: products and platforms, including products like Merkury, GenCX and Dentsu Connect; analytics and insights, spanning analytics, attribution and data engineering; and global services, which includes Dentsu’s creative, media and customer experience management businesses.

Zelcer previously spent 18 years at Merkle, Dentsu’s performance marketing and data arm, most recently as global head of analytics and technology. While there, she developed GenCX AI, which uses generative AI to combine first-party and third-party data.

Dentsu plans to tie GenCX to Merkury, its person-based (as opposed to cookie-based) identity resolution platform, according to Zelcer. The holdco will use generative AI to create audience segments and lookalike audiences with greater speed and accuracy, based on clients’ first-party data and Merkury’s third-party data.

Zelcer has been in the data and analytics space for a long time – long enough to see the term “big data” rise in prominence and die away.

“When I started my career 20-plus years ago versus the vast amounts of data that we have at our fingertips now, everything is big data,” she said.

Zelcer spoke with AdExchanger.

In your global role, have you noticed regional differences?

Yes, especially when it comes to data.

EMEA has different laws regarding first-party data and what we can and can’t do from a media and measurement standpoint. For multinational brands, it’s important to understand the nuances in data, identity and measurement and to be consistent, but also take into account the different regulations in each market.

In EMEA, you can’t gather information on an individual. It has to be more segment-based. That changes for first-party versus third-party data. If your customers have opted in and given you the right to use that data, you can mark it differently.

In different areas around the world, there are geographic limitations [to data collection]. It might be that you can go down to the postal-code level to get demographic information, but not to a household level, for example.

The AI Act recently passed in the EU. What’s Dentsu’s position on the act?

We’re scratching the surface of regulations, and the regulation has not caught up with all the capabilities of generative AI yet. Our point of view is: We are going to protect our client data as much as we can and make sure we’re not putting false information into anybody’s hands.

But [apart] from identity and data, the generative AI work we’re doing is international. There’s nothing preventing us from using the models we’re building globally.

Switching gears, how are third-party cookie deprecation conversations going with clients?

The whole premise of Merkury is to do media activation targeting a person rather than a cookie. We’re working with our clients on building out their identity graphs and how they’re going to be able to identify people once there’s no longer cookies.

What are you recommending?

Knowledge is power. Make sure you know what [signal loss] means for the way that you do your business today and what’s going to change, from the way you buy media to how you measure performance.

Create a road map. You don’t have to change everything on day one. Let’s make sure we’re able to have a strong identity. Let’s make sure we’re putting the right tools in place and start migrating some of your data into these tools.

We’re going to see a lot more brands realizing they need to take advantage of their first-party data. Brands that do that in a meaningful way are going to be ahead.

This interview has been edited and condensed.