Register Now

    Salsify At the Whiteboard

    Where Do AI Shopping Agents Get Their Product Data

    Your product data doesn't stop at the digital shelf. It can shape what AI shopping agents see, understand, and recommend.

    Salsify At the Whiteboard - Where Do AI Shopping Agents Get Their Product Data - thumbnail

    The Mendoza Line of Agentic Commerce

    For years, brands have focused their digital shelf efforts on the products that matter most, while long-tail products often went untouched. But as AI shopping agents read and evaluate product information across the web, outdated or inconsistent PDP content can take on a new significance.

    In this episode of At the Whiteboard, Rob Gonzalez of Salsify explores how product data across the digital, physical, and agentic shelves increasingly influences the shopping experience. Using the “Mendoza Line” as a starting point, he explains why the standard for maintaining product content may be changing, and what brands can do to keep every product working in their favor.

    Hi. I'm Rob, and I'm back At The Whiteboard here to talk about how data and performance across the agentic shelf, the digital shelf, the physical shelf reinforce each other.
    And I'm gonna start with a story here. In the ‘70s, there was a shortstop named Mendoza who was absolutely brilliant with his glove, great defensive player. But he was just terrible at the plate. His batting average hovered around two hundred and people in baseball came up with this idea called the Mendoza line.
    The Mendoza line is how good you have to be at the bare minimum at batting in order to justify being a great defensive player. So he was batting around two hundred and they kept him on the field. But if he was batting one eighty five or something like that, it doesn't matter how good of a defensive player he is. He's getting cut.
    He's not actually playing the ball. And this analogy is great. It works in lots of different places. It's like, where's the bar that you have to get over in order to be effective?
    Now, if you are working on the digital shelf and you're selling on Amazon, you've got maybe your top sellers, maybe it's your top-20 products, your top-30 products on Amazon that are absolutely squared away. Same with Walmart, same with Kroger, same with Home Depot. Any of your major channels, your top sellers, the product detail pages I'm willing to bet are perfect, as perfect as you can make them. But there's a lot of products that are below the line.
    IPG recently did a study of CPG executives that showed that 90% of product detail pages have not been touched in months or years. These are products that were set up at some point, they're not a top seller, and they're just sitting.
    And what that means is that over time, the data on these pages becomes dated. It becomes out of alignment with the rest of your brand, and it becomes out of alignment with the top sellers.
    The agentic shelf is a new issue. On the digital shelf, it didn't matter. Each product kinda stood on its own. You know, if you've got a top-seller on Amazon and it's got 5-star reviews, and it's just making money. It doesn't matter if you've got a 2-star review product that's not really moving that much product and hasn't been touched and is unoptimized. Each product kinda stood on its own. The problem in the agentic shelf here, which is new, is that the agents read everything.
    They'll read every bit of information on your best-selling products and on your crappy products. They read information on your product detail pages that have been updated recently and the ones that haven't been updated recently. So if you've got old products or products that have inconsistent brand messaging or brand messaging that's dated and so on and so forth, they're creating potentially logical inconsistencies. And one of the things that agents are great at is spotting logical inconsistencies in the shopping journey.
    And when they do that, they will tend to recommend other things or they'll say, I recommend this, but with a caveat that this brand has a few products that had these problems and, you know, those types of things. So for you, historically, on the digital shelf, it might have been fine if ninety percent of your products were below the Mendoza line. They're just hanging out there. They get an occasional sale.
    It's all good to go. On the agentic shelf, it's no longer okay.
    So one of the things that we are learning as we experiment with agentic commerce is that everything here, all the data from the digital shelf impacts the agentic shelf.
    This isn't new news by the way. We know the hard way from decades of data at this point that online performance impacts offline performance, and offline performance impacts online performance, and brand marketing impacts. All of these things were sort of related. But now we're actually getting data that shows that product detail pages, especially the long tail, can impact recommendations on the agentic shelf.
    And I'm willing to bet that we're gonna start seeing the agentic shelf impacting the digital shelf. So the agentic shelf starts making recommendations, that will start bringing up digital shelf search rankings and conversion rates and things like that. So what we have to do here as a company is we've gotta make sure that every single product detail page on the digital shelf is optimized. Because if it's not, you've got a new liability as we move into agentic commerce.
    This is possible now using AI. The reason that you would only optimize the top, you know, whatever, 20 products on Amazon is because there's only so much time in the day. You don't really have the time and the resources to touch every single product detail page on every single retailer or make everything perfect. Nobody does. But if you use agents on your behalf internally to do the optimization, you have a fighting chance. So step one to winning on the agentic shelf is to be brilliant at the basics on the digital shelf to make sure every single PDP, every single brand message, every single piece of information about your products everywhere in the internet is correct, is up to date, and is optimized. Thank you.