Subscribe
Share
Share
Embed
10 minutes of expert insights every weekday. Your morning ritual for staying ahead in retail media.
Ep 3 - Ecommerce signals are shrinking. Why retailers must earn new ones
===
[00:00:00] Kiri: Retail media was built on browsing, but if discovery starts to happen upstream inside LLMs and the customer journey on a retailer website gets shorter. The data exhaust that [00:00:15] retailers have relied on for retail media starts to shrink.
[00:00:19]
[00:00:20] Kiri: This expert series is sponsored by Miracle, and and I'm joined again by Amelia VanCamp, head of Agen Commerce at Miracle, to talk about how [00:00:30] retailers earn new signals through experiences that people actually want to engage with. So Amelia, I have a bit of a cheeky question. Um. Why do so many assistance on ai, assistance on retail [00:00:45] websites fail even when the tech itself works?
[00:00:49] I've just had, I've had more worse experiences than good ones. I'd say overall so far.
[00:00:55] Amelia Van Camp: Yeah, I think that's probably a symptom of us being inside of this first wave. [00:01:00] So it's a bit of trial by error. Uh, but then I think beyond that. There is an acceleration inside of magenta commerce where I'm sure all of us feel that you have to do something. And [00:01:15] so in many of these cases, when a retailer has decided to launch a shopping agent, the concept is certainly there, right?
[00:01:24] The idea of how it would work or how it could look is there, but the nuances [00:01:30] between how it works. What's behind the scenes can perhaps get a bit jumbled and that can actually impact in your case, the user experience. So I'm gonna keep saying this for everything in Agentic until it becomes a bit more concrete where [00:01:45] we have some some market standards.
[00:01:48] It's really having a plan in place, not just an idea. The ideas are plenti fold. Here launching a shopping agent. That can be an idea that you have, but again, thinking back [00:02:00] on how you want that interaction with that customer to happen, how you want them to use that shopping agent. Is it specifically for product purchasing?
[00:02:09] Is it also for tutorials? Is it for wider learning? Whatever that may be. Thinking about [00:02:15] all the. Interlocking that needs to happen to enable the best outcome for the user so that they keep coming back. And it's a very touchy topic as well. You wanna go fast, you wanna do something [00:02:30] now because there's pressure in the market, how do you make sure that you do it right?
[00:02:34] And that's not always the easiest thing.
[00:02:36] Kiri: Yeah, that is such a great point. I was at NRF earlier this year and, and so many retailers getting up on stage. We're doing [00:02:45] on, we've got our onsite assistant and this is, wow. This is, like everyone's. So advanced and then I'd go and check out some of these capabilities and every, you know, I think there is a eyes, eyes are bigger than our stomach kind [00:03:00] of situation with some of this technology.
[00:03:02] Where is this? Capability that you're offering with an assistant. Even something that I, as a consumer want. Do I want a recipe in this shopping app? Do I want, don't know. There's, there's a lot [00:03:15] of over, over-engineered things, to your point where the underlying data sources and the way things are working together may not really be robust enough to support those ambitions.
[00:03:28] Amelia Van Camp: Yeah. I think it's [00:03:30] also as well take the history that you have, take the understanding of your, your customer base that you have today, and bring that into your strategy for your onsite experience. Whether that includes the conversational commerce [00:03:45] aspect to it or not, and because of the urgency around this space.
[00:03:51] It's, it's no one's fault and it's completely understandable, but I do see people more often than not, kind of throwing the baby out with the bath water and [00:04:00] saying The way that we've done it is not the way that we're going to do it. There's again, a a ton of information that you already have about your customers today, how they search your site, what PDPs they stay on, what categories they interact with best.
[00:04:14] If you [00:04:15] have tutorials, if you have reviews, how they're interacting with those. Starting to look through that information and figure out how you can synchronize that back into an onsite strategy for ag. Agentic is critical. Otherwise, [00:04:30] you're starting from scratch. When you don't need to.
[00:04:33] Kiri: Something I've been thinking a little bit more about o of a retailer's essential role in this ecosystem, which is product truth, and legitimate [00:04:45] information about the product, claims, pricing, inventory, there's essential information. let's talk a little bit more about that and how. Much of an imperative it is for retailers to get that aspect right, both for [00:05:00] themselves and for the, the, the LLMs that they're interacting with.
[00:05:04] Amelia Van Camp: Yeah. I think to be honest, um, I joke a lot my, when friends and family are like, I don't understand what your job is. And like, it's basically shopping in a modern world. [00:05:15] And when you, when you think about The concepts that already exist, kind of the prerequisites that already exist for retailers, for brands to be able to participate in e-commerce, full stop.
[00:05:27] Those elements are still [00:05:30] very, very, very important inside of agen commerce. I think where the, the waters get a bit murky is when people think about agentic commerce, they start thinking about ai and then we all tend to go off on tangents when it's AI [00:05:45] related because it's this very big space. It's very new.
[00:05:47] There's several different arenas that you can take it in, and at the end of the day, what, what you have to think about is what matters most to your customer and how do you retain that customer for a [00:06:00] long period of time. How do you keep them coming back? The things that you've done in the past for e-commerce are still very critical in agentic today to keep customers coming back.
[00:06:11] If you have pricing anomalies, if you have [00:06:15] shipping delays, if you have a poor after sale. Process. Those are all things that would impact you in e-commerce. Full stop. They're going to impact you the same way negatively inside of Ag agentic where it can [00:06:30] potentially have a double negative impact in Ag agentic is this notion of AI agents.
[00:06:36] Needing to trust you as a retailer and as a brand. And if you have several faults or several negatives [00:06:45] where you haven't been able to fulfill an order, there has been negative reviews based on after sales, negative reviews based on. Purchase experience. Those are things that are going to impact the agent's trust of who you [00:07:00] are and, and who your products are to potentially impact things like discoverability and ranking, um, and even your, your brand showing up inside of a search result.
[00:07:10] Kiri: And that has, when you think about a, a retailer's, [00:07:15] Constituents. The brands who are the suppliers to the retailer, they are really incentivized also to have correct product information. I speak with brands who are, who are asking, what can we do? [00:07:30] To make sure that, you know, our retailers that we're selling to, that they are producing all of this accurate information that the LLMs want.
[00:07:38] 'cause this affects us too, as the brand. um, I think we're gonna get into that a little bit more in the [00:07:45] next episode. what brands can do to participate as well, Amelia, But I think I take away from this is that the signals don't magically appear around consumer behavior. Retailers have to design moments [00:08:00] that earn engagement, the post purchase, the education, the reviews, and in the next episode of this expert series, with Miracle, we'll tackle the big debate around AI ads, where they show up [00:08:15] how trust is protected, and what homework, retailers and brands should do now so that they're ready for whatever new formats emerge.
[00:08:23]