PepsiCo Beverages on testing and learning its strategy in a fragmented media landscape

That fact is one of the things that keeps Katie Haniffy, senior director of media strategy and investment for PepsiCo Beverages, up at night, she said.

“The process has changed,” Haniffy said on the most recent episode of the Digiday Podcast. “Really, there’s probably three things that are going, ultimately, to impact our process.” That’s the booming creator economy, the rise of CTV and retail media. 

On this episode of the Digiday Podcast, we caught up with Haniffy to talk about changes in the media landscape, how PepsiCo Beverages is keeping up with them and, of course, what measurement looks like in a fractured marketplace. 

Below are highlights from the conversation, which have been lightly edited and condensed for clarity.

In-house data amidst privacy regulations 

We have to make sure that we are 100% privacy compliant — keeping the consumer privacy, the data all on the uppity up. That is number one priority. Interesting enough, we have been on this data journey over the last couple of years in preparation of basically everything transitioning to over some sort of digital ad tech platform. We have a great in-house solution. We call it CDNA (Consumer DNA) which is a combination of first-party and third-party data, which includes shopper behavior all the way to behavioral insights, passion points. We can scale that data up and down. That data is really used to inform how the brand and creative is going to connect with the consumer at the planning stage all the way down to the digital activation stage. Being able to utilize this data to activate our media buys has been very very efficient and effective, driving more immediate effectiveness for us to be able to deliver the creative message to the right consumer because we know a little bit more about the consumer. 

The hot topic of AI

In terms of AI, and specifically for media — while there are certainly use cases in creative and other places within the organization — but for media specifically, we’ve been testing AI with a number of partners. OMD, specifically, who’s our agency of record, who is a phenomenal partner on using AI to really optimize our digital buys in real time. So taking some of the manual, hands-on-keyboards off of it, setting the KPI and then letting the machine learning really optimize, find the right consumer at the right time and deliver the message. We’ve seen some early success with it in increasing viewability, increasing sales lift, absolutely. Even there’s some cost efficiencies there. That’s something that we’re going to want to start to expand as the capability grows within the walled gardens and outside the walled gardens. 

Measurement melee 

[Unstandardized measurement] does create more challenges from the traditional mix modeling. We do have our own internal tool called ROI Engine, which is right now, the one source of truth that we have, that really measures holistically, all of our marketing activity — specifically even media. It does get down at the granular level for specific platforms and campaigns so we can then see what is returning better on investment versus something else. That is, right now, the best source we have. Now, we’re certainly going to supplement that measurement with third-party studies, brand lift studies. Going back to retail media, they all have solutions as well that we’ve been partnering with them to do a test and learn strategy to understand what’s driving incremental sales, what is driving brand association.