Data Language Gap Undermining Retail Media Progress and Potential

Reshaping for the Future

While retailers don’t necessarily need to hire more data scientists, they’d be wise to look to the adtech space, said Dave Simon, president of Vibenomics and In-Store Marketplace.   

“There are not nearly enough people who understand how these systems talk to each other in this industry. … If you want a cost-efficient way to go out and steal share and make your products [and] your whole ecosystem function better together, go steal ad tech people from CTV businesses,” he said.  

“CTV is a very, very crowded space, but all of those people know how to take one data set that’s keyed off of a location ID and marry that with a data set that’s keyed off of a cookie ID — that’s what’s required. There are plenty of IT professionals within most of these retail companies who have the chops to stand up a big AWS install that allows them to leverage data. The real key is not generating the insights, but knowing what questions to ask.” 

Epsilon’s Campain echoed this, noting that the biggest value-adds are those who can wear both business and technical hats and ensure data scientists have the info they need for media objectives. 

“You need a business translator who understands the technical world but also understands the business — and who can help each other translate,” said Campain.

There’s good reason to undertake the heavy lifting required of modern data strategies. When looking to the future of in-store retail media, Simon sees a great deal of promise in machine learning. As someone who spent years in mobile app advertising, where deep learning models revolutionized targeting, he’s bullish about its potential to reshape in-store retail media. 

[Related: BP’s Derek Gaskins will talk about their new retail media strategy at P2PI Live in November]

But although retailers hold vast stores of shopper data, their protectionist strategies have them acting more like broadcasters than app developers. 

“There is so much value and intelligence and so much data exhaust being put off when consumers walk into stores. Think about how much time and money is spent on where you put the store, where you put the aisles, how you design the store — all that generates incremental sales value, and they can tie that back to individual users.” 

Unfortunately, the data isn’t currently being used in a sophisticated way to feed machine learning models to improve outcome predictions, he said. When in-store ads are treated like digital impressions and tied to transaction log data, machine learning can reveal correlations that highlight opportunities to influence shopper behavior. 

“It shows not only that you are understanding your own business, but you also understand the mind of the consumer when they’re in your store,” said Simon. 

Progress again hinges upon a mature data strategy. Unlike the mobile app space, where transacting and data collection are straightforward, data sources in retail are incredibly fragmented. 

“Stitching that together is not a small feat,” Simon stressed. “On the retailer’s side, they’re being asked to do things and answer questions for their brands that they may have been asked once a year or a quarter, and now they’re being asked to answer those questions 10,000 times every second.” 

While this optimization is part of many retail media strategies, building the full infrastructure to do it at scale remains a large, top-down and long-term investment for both marketing and IT. 

This article first appeared on the site of sister brand P2PI.