I'm Rob Gonzalez with Salsify. We are back at the whiteboard, and we've got a special session. This is our first-ever guest appearance on the whiteboard. Dave Feinlieb from It's Rapid. Thanks for coming, Dave.
Hey, Rob. Thanks for having me on “At The Whiteboard.”
Yeah. And I'm quite happy that Dave is of a height of me and so is the cameraman. We didn't have to really adjust anything to make this work. So, hopefully, this should go smoothly. Today, we're gonna talk about AI optimization potential on the digital shelf. What problems it's solving, what it can do, and where we're going. So let's dive in.
One thing that most brands struggle with is nailing every single detail for every single product on every single channel across the digital shelf. It's not something that anybody does well. And there's lots of barriers to scale that prevent brands from really succeeding here. On the content and data management side, you've got manual creative production, hard to create retailer-specific versions.
There's some retailers that have a lot of style guidelines around what you can really send them. There's lengthy approval workflows. A ton of the customers that we work with, the descriptions have to be reviewed by lawyers, so that causes delays.
You can't customize every single description for every single retailer, for example. So there's all kinds of data and creative that prevent you from really nailing every product everywhere. But there's also a lot of insights and analytics gaps that most brands are suffering from.
So, Dave, can you explain those?
Sure. Thanks, Rob. Well, one thing we've gotta know is what's going on in the category. So we wanna know what are competitors doing?
What's changing? What are the latest image trends? Maybe the category is really fluid. We wanna make sure we're staying up with those.
We also wanna know, “What are all the gaps in the content?” So imagine if we could create all that content, but we can't figure out what are we missing? Maybe we're supposed to have seven images, but we don't know which three or four do we need. And then what are the differences across retailers?
Maybe this retailer requires one thing, this other retailer requires another thing. Historically, we'd have to go look at all that manually. Maybe we could use some analytics to say, I'm supposed to have seven images, but I only have three. Call that quantitative analytics.
But qualitative analytics is, I wanna look at the entire category. I really wanna understand those images. Are they mobile-optimized? Are they in the right order?
Are they optimized for conversion? Do I have the right titles and copy? Not just the count, but really the content quality. So that's something you can do now that you couldn't do before.
So you got two categories of problems here. Yeah. One is the challenge in just simply getting all your data in place. The second is the challenge of getting signal about that data and then responding to it.
Exactly. So let's talk a little bit more about how AI can change this dynamic. Yeah. You know, so we're in this place today where there's just too much manual steps across the entire cycle.
Yeah. How does AI enable brands to get beyond that?
Well, let's take a step back for a moment.
One thing that's really different now is how we're actually doing the shopping. So I used to go and type in soda or something like that. Now I'm going to the engines, and I'm saying, what's a beverage I could bring to the summer barbecue that's fruity? What's the flavor of that beverage? So the way we're searching is changing. And then the way we're thinking about optimizing the content to match that search is also changing. So those are really two things that are going on with a shopper behavior and then the ability to analyze and optimize that content.
Alright. Let's make this real for folks. Okay? So let's get out of academic land.
Yeah. Let's move to a specific example. Yeah. On one problem, a narrow use case. Yeah.
That people can think about in terms of, alright. I wanna be able to do this across all products. Right? So let's let's go there.
Yeah. Perfect. So we all know that we're on our phones, we're on our mobile devices more than ever before. Probably something like 75, 80% of shopping for ecomm happens on the mobile device.
So, in an ideal world, we'd all have all of our hero images optimized across hundreds or thousands of products. That's the ideal state. The problem is, how do I even know which of my images are optimized and which ones aren't? And what does optimize mean?
So when it comes to analytics, that's sort of subjective. You might say, “Hey, the text should be larger. The contrast should be different.” Things like that.
Those are all subjective measures. But with AI, we can take all those measures, put them into the rule set, if you will, with the AI engine, and say, tell me if my images are mobile optimized or not. I want to know if those images are mobile-optimized according to the GS1 standard. I want to know if they're optimized according to this retailer's standard.
Or maybe my CMO has a particular point of view on optimization. Are they optimized according to our brand standards? So those are all different ways to measure mobile-optimized. Now with AI, we can take that subjective measurement, make it really quantified, really able to measure, and with the click of a button, this is the magic, we can say this is or is not mobile optimized.
Now press the button. We've simplified the pack shot. We've added a callout. It's mobile-optimized. So this is just one? Yeah. Just one.
And the digital shelf is more than just a mobile hero image. Correct. So what are all of the other examples of things that you're gonna have to pay attention to? And today, people are not paying as much attention to because they just don't have the hours. Yeah.
So we wanna make sure our images are mobile-optimized, stopping the scroll, so to speak, easy to read on a mobile device. We also want to look at all the other aspects of our product detail page. So that could be, “Do I have the right images? Are they optimized for conversion? Are they sequenced in the right order? Am I telling the full story that I want to communicate to the shopper so they understand my product and can make a purchase decision? Do I have all the right copy? “
Imagine back to that question, “What am I going to bring to the barbecue?” I want to make sure all that content is optimized. So you've got mobile, you've got image sequencing, you've got conversion, you've got what we call content quality across the whole product.
Yeah. It's just over and over again for images. Yeah. Beyond images.
Beyond images, there's tons of opportunities. So we used to measure things like the character count, the title length, or the number of words in the copy. Now we're doing things like not just keywords, but are we asking and answering the right questions in that content? So we'll show up in the large language models in Sparky and Rufus and engines like that. So you've got all of your imagery, you've got all your copy, you've got above the fold, you've got below the fold, tons of content you can work with. With AI, you can now analyze that, and you can click a button and optimize the entire catalog.
Yeah. One example that we've used on the whiteboard previously is if you go to an Amazon product detail page, Rufus actually tells you what people tend to be asking in the product category for the product detail page. Right. And very often, brands don't even have that copy anywhere in the product detail page, they might not have written it internally.
Exactly. Right? So then you've got this challenge of getting signal from Rufus, finding content gaps, putting it through creative, republishing, going through the whole cycle.
And again, this is just one example of something that you should be doing across all of the details on all the product detail pages.
And I don't know if you, you know, you've looked at your content playbook and you're thinking to yourself, where do I start? And AI is really the answer to knowing, I know what I'm supposed to do in an ideal world. Give me the analytics to tell me all the gaps, all the opportunities, and now let me go do the optimization, not just for a few products, but hundreds or the entire catalog. Yeah.
So, where we're going with this is I know for years and years, the vast majority of brands would honestly, they would, know, set up an item once and publish it, and may not touch it for a while. Right? Best practice is that you're touching it at least quarterly. A lot of brands will do an annual refresh.
Right. But, you know, you're not really optimizing every single product that much. Maybe your top 50 products on Amazon, maybe your top 10 products on Amazon you're looking at more frequently than that. But what you're talking about is getting to this promised future that people have wanted to get to forever, which is a continuous optimization cycle where things are constantly being analyzed, looked at, and improved almost automatically.
In real time. And we're talking every single product and every single page.
Anyway, this has been a dream that people have been talking about for as long as I've been doing ecommerce. It's not been realized. How close are we now given what we're talking about AI's potential on the digital shelf?
It's not as far off as we might think. It's happening today. So we have brands we work with who are doing this continuous optimization loop.
And like you said, maybe they're going from annual to quarterly, but maybe they're going from quarterly to just ongoing. Optimizing the whole catalog. They're looking at moments during the year, seasonal opportunities, things like that, that they've always wanted to get to, but they haven't been able to get to. So yes, you can apply AI today, understand, get the insights, look at all the gaps and the opportunities, and then do the optimization.
That's awesome.
Alright. Well, thanks so much for joining me at the whiteboard, Dave. Thanks, Rob. It's been great to be here. We'll be back soon.
.svg)