“The Sell Sider” is a column written by the sell side of the digital media community.
Today’s Q&A is with Michael Schuh, VP of media strategy Kroger Precision Marketing’s 84.51°.
Kroger, the largest grocery chain in the United States, is capitalizing on major swings in shopping patterns since the COVID-19 pandemic as it expands its retail media business.
For one thing, online and in-app groceries went from being a barely consequential fraction of Kroger’s daily shoppers to a primary source of demand. And the trend is holding. Even compared to the worst period of nationwide lockdowns last year, Kroger’s online ordering continues to grow. Online sales have more than doubled since the end of 2019, according to the grocer’s most recent earnings in June, and the company predicts another doubling by the end of 2023.
On top of this secular trend, Kroger Precision Marketing VP of media strategy Michael Schuh said the online advertising industry has come to realize that “a reset needs to happen” soon with a focus on first-party data as a source for targeting and measurement.
The opportunity for retail media in the new programmatic landscape is hardly new – Amazon, Walmart and Target, among other large retailers, are all investing heavily in ad tech.
AdExchanger caught up with Schuh about how Kroger puts its retail data to use in digital media.
AdExchanger: What media, owned-and-operated or not, is available on the ad platform?
MICHAEL SCHUH: We have a diverse portfolio of products. The one that brands probably think first about, especially in a retail environment, is on-site search or product listing ads within search results on Kroger.com and in the mobile app. We have on-site display as well within the Kroger digital properties.
And then off of the owned properties, we’re also focused on leveraging Kroger data to build audiences that connect into direct publisher relationships. We call it off-site media. One partnership is with Roku; that was formed a little over a year ago. We also can activate Kroger audiences and measurement across social channels, across the open web, for email campaigns, and as I said there are a number of other direct publisher partnerships.
How would you categorize KPM, in terms of being a DSP, SSP, some hybrid or none of the above?
It depends which part of our portfolio they’re working with. To some extent, we’re clearly the supply side. We have inventory on Kroger properties that we are opening up, and we don’t open that up in other DSPs. We buy directly on our media platform and with our assets.
But then we also use our audiences and our measurement products to feed into and influence what a brand learns with their own spend in, say, some other DSPs or another environment.
Our focus is on enabling our audiences and measurements to live wherever people want to buy, meeting advertisers where they are. So yes, we can buy it ourselves. We have our own buying and execution team here. And we can operate sort of as a managed service in that capacity. But we’re also enabling self-serve use cases, wherever brands want to focus their time.
If you’re buying direct inventory, you buy that directly through our self-serve tools. And Kroger data can, with cookies going away and any number of other obstacles for online data, help to make that campaign more accountable and more performance-oriented, even if we don’t manage the spend ourselves.
Are there examples of direct publisher integrations that have had success for you?
I mentioned the partnership with Roku. They are the number one streaming platform in the US. What it means is we can pipe our audiences into their inventory, and connect the actual media exposures and tie that back to sales. That’s also an inventory deal for media that we can sell directly.
We also have a partnership with Meredith. Consider a title like Allrecipes, the no. 1 recipe and grocery focused site. There are the contextual insights that happen on a site like Allrecipes and other Meredith publications, paired with the behavioral insights and deterministic insights that happen in our ecosystem. Those two together are really powerful.
You’ll see us continue to expand our publisher direct relationships over time.
Are there ways to use your data, aside from connecting to transactions and store lift?
The main source that we pull from is our first-party purchase data. Brands want to know what customers bought, but maybe just as important is what they might be most interested in buying next. There are a lot of really interesting and useful signals we can pull out of that data.
One example we use is when an existing brand is launching a new line extension or something with new flavors. We can not only pull their existing buyers and buyers in the category, but who else buys that similar flavor from the rest of the store? So, say it’s a new flavor of chips or something, right? Like they’re testing hot sauce flavored chips. We can pick people who are buying chips and people who buy that brand. But also people who buy hot sauces, or things that typically are in the cart with hot sauce. There are some really unique combinations that we can create.
There are also other contextual signals and product attributes that could be important. Does this person go for deals like, “Buy one get one half off”? And there are other ways to target by price perception – like if they’re likely to choose the product that’s on sale, or is the cheapest on the shelf. We actually use thousands of different data signals, as part of the intelligent audience science in our system. And there is custom work off of brand data in a lot of cases that really builds the most optimal audience within our ecosystem.
Is your measurement report deterministic – you reached X number of Kroger shoppers and Y of them purchased your product – or is it a more probabilistic measure of general sales lift?
A little bit of both, but it’s more so the former.
We report back on sales lift, on household penetration lift and on return on ad spend. That doesn’t mean we don’t also report on some of the standard intermediate metrics on impressions, on viewability and on click-through rate, but we know that a brand is focused on real outcomes. Brands want to see what worked and what drove sales, incremental growth and new customers. So that’s what most of our reporting focuses on.
What’s the rate of sponsored vs organic listings for searches on the site or app?
The exact rate is a tricky question to answer because searches on Kroger actually leverage a lot of that customer data and are personalized if there’s any shopper history. And we really believe that that’s an important part of the experience.
Take a search for “granola bar.” If there’s one particular brand or kind of granola bar that someone purchases all the time, we at Kroger should make it really easy for them to find that product in the number one spot or certainly above the fold.
That’s not the way it works on all retail media platforms. There can be an inverse dynamic where, an advertiser can pay to get in front of that and suppress the product that customer wants, if another brand pays enough. We can leverage those signals, but for us the right thing to do for a customer is not to bury their favorite products when they’re searching online. So we try to make it easy for them.
That lets brands have influence when customers are actually looking for product research or discovery. Eighty-five percent of the top 500 search keywords on Kroger.com are unbranded. People are mostly searching for things like “granola bars” and not searching for things like “Quaker granola bar.” We want the advertiser bidding for an impression when someone is considering what to buy, though. Not when they know exactly what they’re going to buy, as a chance to get in the way of that.