WEBVTT

00:00.051 --> 00:03.894
[SPEAKER_02]: Before a Gentic search, what would happen when somebody wanted to find your product?

00:03.934 --> 00:04.615
[SPEAKER_02]: They would Google.

00:04.775 --> 00:07.417
[SPEAKER_02]: They would browse through the first second, third page results.

00:07.797 --> 00:10.379
[SPEAKER_02]: And you know, you hoped you got them as a customer there.

00:10.439 --> 00:13.041
[SPEAKER_02]: Otherwise, they pressed the backbone on their browser and went to the next result.

00:13.502 --> 00:26.232
[SPEAKER_02]: Now, with a Gentic search, the user still searches for what they want, but the AI grabs the first 10, 20, 30 results, compiles a recommendation, and then responds to the user instead.

00:26.912 --> 00:37.922
[SPEAKER_02]: So the difference here is that you don't actually get the user on your page and now you have to optimize for what the agent is going to say about your product and whether or not it's going to say anything at all.

00:50.584 --> 00:52.105
[SPEAKER_00]: Hello and welcome.

00:52.205 --> 00:56.227
[SPEAKER_00]: I'm Chloe Thomas the host of this ecommerce marketing focus podcast.

00:56.667 --> 01:03.730
[SPEAKER_00]: If you're not familiar with our format, well, we release five episode mini series that dive deep into one ecommerce marketing topic.

01:04.151 --> 01:10.674
[SPEAKER_00]: That could be channels like email or SEO or Facebook ads or a big topic like acquisition, retention or analytics.

01:11.474 --> 01:19.943
[SPEAKER_00]: In each episode, new ones released on Wednesdays, I get to interview a different expert to explore the latest device on making that marketing method work for you.

01:20.443 --> 01:26.829
[SPEAKER_00]: And then after each five episode, many series we take a week off to give you time to take action on what you've learned.

01:27.570 --> 01:31.315
[SPEAKER_00]: At the moment, we're spending five episodes exploring AI search.

01:31.655 --> 01:37.222
[SPEAKER_00]: Yes, our first series completely dedicated to getting your products recommended by AI.

01:37.262 --> 01:45.512
[SPEAKER_00]: We did a mixed up series earlier this year with SEO involved as well, but this time we'll all about AI search and we're going to be exploring it from a number of different angles.

01:46.173 --> 01:48.234
[SPEAKER_00]: But first of, we have to cover the basics.

01:48.254 --> 01:52.697
[SPEAKER_00]: We have to cover the key stuff of how you get your products recommended by AI.

01:53.097 --> 01:56.319
[SPEAKER_00]: And that's what our first guest in the series is joining us to explore.

01:56.699 --> 01:57.960
[SPEAKER_00]: He's a mine of information.

01:58.500 --> 02:03.143
[SPEAKER_00]: You're going to hear some surprising insights here and some stuff you won't probably haven't heard of before.

02:03.883 --> 02:05.504
[SPEAKER_00]: Plenty of to-do's coming out of this.

02:05.584 --> 02:11.028
[SPEAKER_00]: Get prepared for the AT20 of it all and the key three areas you need to address.

02:17.251 --> 02:19.412
[SPEAKER_00]: In this episode, I'm chatting with A.I.

02:19.452 --> 02:21.073
[SPEAKER_00]: Sirt expert J.L.

02:21.113 --> 02:21.554
[SPEAKER_00]: Carke.

02:22.234 --> 02:27.637
[SPEAKER_00]: J. co-founded the most popular loyalty software in the world called Smile while he was in college.

02:28.057 --> 02:35.221
[SPEAKER_00]: After that, he founded Ferra, one of the most popular review apps for Shopify merchants, and now he's co-founded Ocelo.

02:35.582 --> 02:46.428
[SPEAKER_00]: A marketplace listing platform optimized for a gentic workplace, used by Shopify fashion brands, listing on Nordstrom, Macy's Blooming Dales and Target Plus, hello, J.

02:47.127 --> 02:47.668
[SPEAKER_02]: Hello.

02:48.388 --> 02:50.531
[SPEAKER_00]: Very cool to have you here on the show.

02:50.551 --> 02:52.313
[SPEAKER_00]: Well, it's always cool to catch up with you.

02:52.353 --> 02:55.036
[SPEAKER_00]: And the fact we get to record at this time is just extra special.

02:55.577 --> 03:01.203
[SPEAKER_00]: Not least because you are going to be talking about a topic that you are deep in in many ways.

03:01.524 --> 03:03.946
[SPEAKER_00]: So very interested to hear your take on all of this.

03:04.367 --> 03:07.490
[SPEAKER_00]: How did you get obsessed with a high set?

03:09.057 --> 03:17.061
[SPEAKER_02]: think last year, the light bulb kind of, maybe two years ago, actually.

03:17.101 --> 03:23.404
[SPEAKER_02]: The 22 years ago, the light bulb really clicked for me on the possibilities with agentic things.

03:23.924 --> 03:32.968
[SPEAKER_02]: It was kind of a switch in the ecosystem where people, we always had this AI idea where AI can convert, it can parse things that are unstructured, but it was a first time

03:37.070 --> 03:41.212
[SPEAKER_02]: work on your behalf and make decisions and make subsequent decisions.

03:41.992 --> 03:56.498
[SPEAKER_02]: And after that, I fell in love with this idea of having this separated brain that could do things for you and started working on Ocelo before and at that time two years ago, you know, just so that people understand how new

04:03.000 --> 04:03.760
[SPEAKER_02]: floated out there.

04:03.800 --> 04:09.384
[SPEAKER_02]: There were some people that were putting it out as maybe this is what's going to happen, but it was really expensive to run.

04:09.424 --> 04:11.625
[SPEAKER_02]: It would cost a certain amount and it was really slow.

04:11.665 --> 04:16.448
[SPEAKER_02]: It never made sense to do a checkout at all ever with two years ago.

04:17.149 --> 04:24.493
[SPEAKER_02]: Two years later, it makes a lot of sense, especially anthropic yesterday, at least on at 5.5, which is

04:31.792 --> 04:34.974
[SPEAKER_00]: Yeah, it's kind of, it's the obsession that never stops giving.

04:35.655 --> 04:53.088
[SPEAKER_00]: It isn't always, there's always something going on, but you're here to take us through how to get AI to recommend your products, which there's a lot of people talking about, but I think as everyone's about to discover, your take on it is somewhat more data backed and transparent than many as we go into this.

04:53.108 --> 04:56.771
[SPEAKER_00]: So how do you get AI to recommend your products, Jay?

04:57.459 --> 05:12.202
[SPEAKER_02]: Yeah, I mean, that really is the question that every merchant right now seems to be asking is, how do I show up in the recommendation now that the search engine optimization matters a lot less.

05:12.782 --> 05:22.884
[SPEAKER_02]: And you mentioned that that focusing on data back to recommendations, well, the challenge here is that because every every entrepreneur out there knows that

05:26.785 --> 05:54.322
[SPEAKER_02]: everybody wants a silver bullet and so there's a lot of products up there right now that are being marketed as oh we'll solve your we can get you listed we can get you recommended we can monitor this and a lot of it is just crafting their own product and they're just hoping that you'll use it attach to it and then once gives them time you know for maybe a few years to figure it out and by then you'll already be a customer so there are very few resources where

05:55.483 --> 06:03.007
[SPEAKER_02]: we can actually gather data and understand what's happening and look at trends without there being such a bias.

06:03.507 --> 06:11.332
[SPEAKER_02]: And for Ocelo, we automate the marketplace listings using AI agents, but we are focused on the work for listing.

06:11.872 --> 06:19.636
[SPEAKER_02]: And so, you know, agentic optimization is really important to all of our clients, but it's not a reason to use our app.

06:19.756 --> 06:22.017
[SPEAKER_02]: So I like to hope I'm less

06:23.562 --> 06:33.145
[SPEAKER_00]: If you do kind of fit in this interesting space if you've come on to talk about something that you're not actually selling, which I think makes you almost a completely unique guest on this show.

06:33.605 --> 06:37.906
[SPEAKER_00]: So you shared some notes with me in advance of us recording this.

06:38.026 --> 06:49.370
[SPEAKER_00]: And one of the really interesting things you say in that is that 80% of getting AI to recommend your products is stuff you're already doing, which suggests is not that much for us to do.

06:50.682 --> 07:06.909
[SPEAKER_02]: Yes, yes, after I decided that I really do spend more time understanding what's going on, which was mostly driven by our sellers that use Ocelo and previous merchants that I have remained friends with that we're all telling me that, hey, can you help with this or do you have any idea?

07:07.409 --> 07:08.710
[SPEAKER_02]: So I went out to try to figure out.

07:09.270 --> 07:12.691
[SPEAKER_02]: you know, what data can I gather and what, what can we trust?

07:13.171 --> 07:29.196
[SPEAKER_02]: And one of the most valuable resources that I brought back so many conclusions out of was Kate Rigody and Kyle Rizley's talk that they gave at the shopfly.dev conference a few months ago about agentic optimization.

07:29.796 --> 07:34.098
[SPEAKER_02]: And they provided a lot of data that gave us some insights on what

07:34.878 --> 07:42.684
[SPEAKER_02]: what merchants should be doing and what partners should be doing in order to help merchants prepare for what's going on with agentic search.

07:43.304 --> 08:00.937
[SPEAKER_02]: The conclusion was basically 80% you should be doing this anyways and the other 20% was we think this is what you should be doing and that's based on their research and we say we think because it's changing so fast so far the research that they're running which is across all their sellers that's what they came up with.

08:01.437 --> 08:12.587
[SPEAKER_02]: And maybe I'll just take a step back here and like mention, you know, what is agent tick search like what what is it, you know, maybe people are don't fully understand why it's such a big deal.

08:13.067 --> 08:19.132
[SPEAKER_02]: Well, before agent tick search, what would happen when somebody wanted to find your product, they would Google.

08:19.833 --> 08:26.755
[SPEAKER_02]: or search on a search engine, they would browse through the first second third page results sometimes, usually just the first, honestly.

08:27.435 --> 08:29.416
[SPEAKER_02]: And if they found your part, they'd click on it.

08:29.556 --> 08:34.217
[SPEAKER_02]: Then you had an opportunity to sell them on why they should be staying there.

08:34.837 --> 08:38.598
[SPEAKER_02]: And then, you know, you hoped you got them as a customer there.

08:38.678 --> 08:41.539
[SPEAKER_02]: Otherwise, they pressed the backbone on their browser and went to the next results.

08:42.547 --> 09:01.934
[SPEAKER_02]: Now with a gentic search, the trend which appears to be now all the experts agree this is almost definitely the new way of search that the old ways pretty much dying is that the user still searches for what they want, but the AI grabs the first 10 20 30 results.

09:03.997 --> 09:07.821
[SPEAKER_02]: Compiles a recommendation and then responds to the user instead.

09:08.482 --> 09:20.314
[SPEAKER_02]: So the difference here is that you don't actually get the user on your page and now you have to optimize for what the agent is going to say about your product and whether or not it's going to say anything at all.

09:20.955 --> 09:45.837
[SPEAKER_02]: That's why everybody's like, you know, pulling their hair out, realizing that, oh, my god, all these, you know, a decade of SEO, you know, did I just throw that all that out, all my accumulated SEO and yeah, so based on that, what we know is the following, we know that fewer humans are visiting your P pages, your product, this detail pages and you probably already noticed that in your stats, the tracking is going down, although,

09:46.958 --> 10:00.463
[SPEAKER_02]: sidebar here, I will mention that Google did recently make an update to how they track to start including more agentic bots in your search data.

10:00.483 --> 10:07.685
[SPEAKER_02]: So you may have also noticed, think last week or something, two weeks ago, that there was another bump up in traffic, and that's probably what it was.

10:07.825 --> 10:12.967
[SPEAKER_02]: They changed it was the num attribute in the fields for any SEO experts

10:16.648 --> 10:23.474
[SPEAKER_02]: So, but we do know still, if you are humans, because those are still bots, even though it shows the thing, they are visiting your product page and you have to expect that.

10:23.935 --> 10:27.077
[SPEAKER_02]: We know that people are trusting AI recommendations more and more.

10:27.518 --> 10:31.962
[SPEAKER_02]: So, as the AI's assistants are getting smarter with you, they're trusting them more.

10:32.642 --> 10:33.703
[SPEAKER_02]: And we know that,

10:34.884 --> 10:36.425
[SPEAKER_02]: AI agents act like humans.

10:36.945 --> 10:38.726
[SPEAKER_02]: They are optimized to act like humans.

10:38.826 --> 10:40.066
[SPEAKER_02]: That's the whole point.

10:40.427 --> 10:43.248
[SPEAKER_02]: So it is very different than SEO.

10:43.528 --> 10:44.869
[SPEAKER_02]: Search is not optimized.

10:45.349 --> 10:47.130
[SPEAKER_02]: SEO is not optimized to act like humans.

10:47.150 --> 10:54.073
[SPEAKER_02]: It's optimized to act like a code, if statements that will branch and find the right thing.

10:54.473 --> 10:57.454
[SPEAKER_02]: AI agents are optimized to think and behave like humans.

10:57.955 --> 10:58.615
[SPEAKER_02]: Those are the three things.

10:58.635 --> 11:01.436
[SPEAKER_02]: So fewer humans are going to visit your product detail page.

11:01.876 --> 11:02.817
[SPEAKER_02]: That's what you need to think about.

11:03.337 --> 11:04.778
[SPEAKER_02]: And that's what we know is happening.

11:05.578 --> 11:07.699
[SPEAKER_02]: More people are trusting the AI's recommendations.

11:07.719 --> 11:09.379
[SPEAKER_02]: So the AI says you're the best option.

11:09.980 --> 11:12.241
[SPEAKER_02]: You're probably, they're going to purchase a lot more.

11:12.841 --> 11:18.623
[SPEAKER_02]: And agents are optimized for human interaction, not for computer to computer interaction.

11:19.023 --> 11:22.645
[SPEAKER_02]: So that'll make more sense when I went with the 20%.

11:23.045 --> 11:26.686
[SPEAKER_02]: And so that talk came down to 80% of the things.

11:26.706 --> 11:28.427
[SPEAKER_02]: So now how do you get the AI to recommend?

11:28.507 --> 11:29.788
[SPEAKER_02]: 80% same things you were doing.

11:30.568 --> 11:33.750
[SPEAKER_02]: You should have been doing anyways, and then the 20% will break down here.

11:34.251 --> 11:34.551
[SPEAKER_00]: Excellent.

11:34.571 --> 11:36.552
[SPEAKER_00]: Let's get into the 20% them.

11:36.572 --> 11:40.315
[SPEAKER_00]: What's the first group of new things we should do?

11:40.355 --> 11:42.076
[SPEAKER_00]: The specificity bits.

11:43.137 --> 11:43.457
[SPEAKER_02]: Yes.

11:43.657 --> 11:50.342
[SPEAKER_02]: So the 20% go up breaks down to three things, specificity, social proof, and taste.

11:50.982 --> 11:57.607
[SPEAKER_02]: Those were the three things that in the dot-dev Shopify talk that Kate and Kyle gave,

11:58.187 --> 12:10.618
[SPEAKER_02]: they effectively broke down to those are the things that were in the 20% that they appeared to make a difference of whether or not AI recommended your product over the other options when it was included in that top set.

12:10.978 --> 12:13.641
[SPEAKER_02]: So the first thing is specificity.

12:13.701 --> 12:16.223
[SPEAKER_00]: They really should have picked an easier word, shouldn't it?

12:16.243 --> 12:21.708
[SPEAKER_00]: It's hard to read, it's hard to say, I'm not sure it's a real word, but yeah, specificity.

12:22.328 --> 12:23.950
[SPEAKER_00]: What do we do for specificity?

12:25.575 --> 12:35.026
[SPEAKER_02]: So what that means is that it means the more specific your listing is, the more likely to be recommended to the AI.

12:35.086 --> 12:36.107
[SPEAKER_02]: And that makes sense.

12:36.628 --> 12:45.337
[SPEAKER_02]: The customer is looking for a red shirt and your product detail page just says shirt, and doesn't indicate that it's available in red.

12:45.858 --> 12:52.481
[SPEAKER_02]: that page is not going to be included in the recommended top three that the AI chooses, even if it sees your product page.

12:52.881 --> 13:00.804
[SPEAKER_02]: Even if you are fourth result on our number one result on Google, you know, when it's Googling on on the customers, we have it's not going to include you.

13:01.324 --> 13:08.467
[SPEAKER_02]: Now Shopify, and I probably other platforms are doing things to optimize to make sure that the agents

13:09.447 --> 13:13.868
[SPEAKER_02]: know that that product is available in different colors, but I won't get into the details of that.

13:13.908 --> 13:17.089
[SPEAKER_02]: And for all intensive purposes, we won't complicate with that.

13:17.649 --> 13:25.272
[SPEAKER_02]: What you need to know, as a seller, would be that you need to ensure your titles are specific.

13:25.932 --> 13:27.172
[SPEAKER_02]: If there is something about the

13:33.637 --> 13:38.278
[SPEAKER_02]: likely matched will be what the customer is searching for included in title.

13:38.819 --> 13:39.399
[SPEAKER_02]: Easy as that.

13:40.019 --> 13:50.142
[SPEAKER_02]: If it's available in multiple colors, don't include one color in the title, but just make sure your product detail page is merged in ones that the colors are available on the cut when the agent visits it.

13:50.602 --> 13:52.943
[SPEAKER_02]: But the most important thing is specificity.

13:53.263 --> 13:57.967
[SPEAKER_02]: The second thing is the Google rank of number one is probably no longer worth your effort.

13:58.247 --> 14:10.676
[SPEAKER_02]: Being number one versus being number three on a Google search, the effort level is, you know, it's a diminishing rate of return and that has with the AI search has diminished a lot more.

14:11.197 --> 14:15.500
[SPEAKER_02]: So the end result is that if let's say that you are a

14:16.821 --> 14:21.525
[SPEAKER_02]: In our case, you know, Ocello's a marketplace listing platform optimized for fashion brands.

14:22.586 --> 14:41.842
[SPEAKER_02]: One thing we've seen a lot of success with is that focusing on fashion brands and making that part of our narrative and not worrying about ranking on the marketplace listing platform search has benefited us a lot and being recommended by AI and believe it or not even just being around for

14:42.482 --> 14:46.425
[SPEAKER_02]: You know, our websites only are actually live for a few months.

14:47.026 --> 14:50.969
[SPEAKER_02]: And yet, we've already had somebody tell us that AI recommended our solution to them.

14:51.069 --> 14:55.772
[SPEAKER_02]: And they were, they ended up using our platform and they're very happy because they were a fashion brand.

14:56.233 --> 14:58.394
[SPEAKER_02]: The truth is, our platform works for non-fashion brands.

14:58.875 --> 15:09.403
[SPEAKER_02]: But it is better to be on the first page and be focused, be specific than it is to be on the first result and not specific.

15:09.923 --> 15:16.371
[SPEAKER_02]: because the AI is going to always choose the more specific thing to what the user asked.

15:16.872 --> 15:19.255
[SPEAKER_02]: And users aren't going to ask for the general things.

15:19.716 --> 15:23.501
[SPEAKER_02]: The browsing is now being done with the AI and the AI is going to clarify with the customer.

15:24.202 --> 15:36.847
[SPEAKER_00]: Yeah, it feels like with all of this, this specificity bit Jay that it's kind of like long tail keywords, you know, like we used to have top top top key short keywords and long tail keywords.

15:36.867 --> 15:43.010
[SPEAKER_00]: It's like the only strategy now is to go long tail to be specific about your title tags.

15:43.050 --> 15:51.954
[SPEAKER_00]: If you're selling the red snake skin silk blouse, then that's what you're optimizing for not.

15:52.654 --> 15:57.998
[SPEAKER_00]: flowers in high notes, nakeskin blouse, not silks nakeskin blouse, but the whole shabang that.

15:58.619 --> 15:59.219
[SPEAKER_02]: Exactly.

15:59.259 --> 16:10.628
[SPEAKER_02]: And if you do the math, it makes a lot of sense, because previously when you were optimizing, you would say, okay, if I go snakeskin blouse, then I'm getting four hits per bleaker something on it.

16:10.968 --> 16:17.133
[SPEAKER_02]: Like, you know, you do the math or at one percent or two percent conversion rate if you're like incredible, you know, and you're basically at zero.

16:22.157 --> 16:25.579
[SPEAKER_02]: those four hits the conversion rate is skyrocketed higher.

16:25.920 --> 16:28.681
[SPEAKER_02]: It is up to like now 20% conversion.

16:29.082 --> 16:40.909
[SPEAKER_02]: So do you go for 20% conversion on four hits or three percent conversion on, you know, on 20 hits and it's still going to be right to the math theory.

16:41.349 --> 16:41.730
[SPEAKER_00]: Something.

16:42.375 --> 16:45.656
[SPEAKER_02]: Yeah, it's definitely now worth it to do the long tail.

16:45.716 --> 16:47.216
[SPEAKER_02]: So you're exactly right.

16:47.476 --> 16:50.437
[SPEAKER_02]: Long tail is now, and that's what we'll see with the long tail.

16:50.457 --> 17:00.938
[SPEAKER_02]: If you were looking at it as like this graph like that, the long tail is increasing, and then the heavy is like, you know, it's balancing out, you know, which is interesting.

17:01.539 --> 17:07.320
[SPEAKER_00]: So let's go on to the second part, the social proof part of the new 20% of things we have to do.

17:07.440 --> 17:08.940
[SPEAKER_00]: What does that entail?

17:10.161 --> 17:29.770
[SPEAKER_02]: actually one one thing I just want to add to this specific city don't spend too much time trying to augment and like there's this whole whole world right now people saying all you need to enrich your product data product detail pages with these attributes and like sort the data and like sort the attributes.

17:30.250 --> 17:31.151
[SPEAKER_02]: please do not do that.

17:31.491 --> 17:39.917
[SPEAKER_02]: The AI is going to be better at figuring out those things than your crappy app that you've installed that's trying to like organize it.

17:40.218 --> 17:46.883
[SPEAKER_02]: The only thing that you can do that will help right now and this is only right now is describe the image is better.

17:47.183 --> 17:53.348
[SPEAKER_02]: So if for example, if your shirt is lined or unlined, just add that in the description that it's lined or unlined.

17:53.488 --> 17:58.571
[SPEAKER_02]: That's where we find with our fashion brands, they see the most benefit of enrichment.

17:59.432 --> 18:07.614
[SPEAKER_02]: Now we're saying, I, my belief is that it's going to be temporary because right now the agentic search is not looking at the images, but soon it will.

18:07.994 --> 18:21.437
[SPEAKER_02]: It's probably only doing that because it's expensive and you know slower, but the models are getting so fast and so, and the costs are just going down so fast that eventually it's just going to look at the image in which case makes no sense to augment anything.

18:21.777 --> 18:23.098
[SPEAKER_02]: Then you just focus on the next two

18:29.207 --> 18:31.112
[SPEAKER_00]: It's a excellent segue to social proof.

18:31.353 --> 18:32.416
[SPEAKER_00]: So what do we do in social proof?

18:32.436 --> 18:36.346
[SPEAKER_00]: Because this is the bit which surprised me the most when you were telling me about it.

18:37.236 --> 18:46.325
[SPEAKER_02]: So just to recap, we said that Shopify's talk at the dot to have and their researchers said that 80% of the stuff you should be doing 20% you need to do a little bit more of.

18:46.725 --> 18:50.309
[SPEAKER_02]: And out of that 20% was specificity, which is what we talked about.

18:50.729 --> 18:52.231
[SPEAKER_02]: The second one is social proof.

18:52.951 --> 19:00.318
[SPEAKER_02]: And what they found was that when real humans verified humans were talking about your product,

19:01.039 --> 19:02.780
[SPEAKER_02]: then the AI ranked that high.

19:03.440 --> 19:21.607
[SPEAKER_02]: There was a stat at one point, which said that the AI recommended recommended products that were mentioned positively on Reddit, a staggering, I can't remember the exact percentage, so I don't want to side it, but it was, it was a staggering, the high rate of recommendation.

19:21.667 --> 19:23.968
[SPEAKER_02]: So you think about what that means.

19:24.008 --> 19:24.929
[SPEAKER_02]: It means that you could get the

19:26.069 --> 19:33.192
[SPEAKER_02]: third page, your product detail page is ugly, you're marketing sucks and some person on Reddit just said, hey, I tried this is really good.

19:33.532 --> 19:34.673
[SPEAKER_02]: And AI was recommending it.

19:34.693 --> 19:39.335
[SPEAKER_02]: You're now beating like the big dogs about it, just because somebody on Reddit said that.

19:39.395 --> 19:46.458
[SPEAKER_02]: So that's like community, investing more in community, third party reviews, investing more into that.

19:46.598 --> 19:48.098
[SPEAKER_02]: The AI is considering that.

19:48.379 --> 19:50.600
[SPEAKER_02]: And investing in third party pages.

19:50.760 --> 19:50.920
[SPEAKER_02]: So

19:51.880 --> 20:18.773
[SPEAKER_02]: This is a little bit, uh, uh, I'm obviously biased on this because Ocelo helps, uh, brands sell on marketplaces like Nordstrom.com and macy.com target, the target plus market place, but we found that listing on those sites when when the AI sees you on a more trusted brand like that that is less likely to lie because obviously I'm more on the line of line, you know, they could, they, there's more to, to get sued for.

20:19.233 --> 20:20.514
[SPEAKER_02]: Then the AI actually is more

20:21.574 --> 20:27.978
[SPEAKER_02]: So there's a value of just getting listed on those, even though you may not see as many sales as you expect.

20:28.438 --> 20:34.422
[SPEAKER_02]: So just to recap here on the social group, third party review sites, very more valuable.

20:34.982 --> 20:49.291
[SPEAKER_02]: On-site reviews, again, I'm biased that I founded one of the most popular with these reviews software, software for Shopify merchants Farah before, but it does appear, it's factoring those in, but not as much as the third party pages.

20:49.671 --> 20:58.755
[SPEAKER_02]: So by far, even more than the retail marketplace listings and the third part of review site, it's community communication.

20:58.835 --> 21:12.240
[SPEAKER_02]: So Reddit, other community boards, even in social media communities, if people are talking about the brand positively, if you can encourage that, that is very valuable, much more than before.

21:12.921 --> 21:13.201
[SPEAKER_00]: Nice.

21:13.421 --> 21:15.302
[SPEAKER_00]: Okay, let's talk on number three, then.

21:15.982 --> 21:18.043
[SPEAKER_00]: Taste, what on earth does taste mean?

21:18.932 --> 21:19.272
[SPEAKER_02]: Yes.

21:19.552 --> 21:34.663
[SPEAKER_02]: So taste, if you have been spending time trying to figure out AI search, AI, or a gentic search optimization, you've probably heard this term before, but for anybody who hasn't, it's very odd.

21:35.303 --> 21:37.745
[SPEAKER_02]: And what it means,

21:38.245 --> 21:57.406
[SPEAKER_02]: You know, why don't I, I'll, I'll just say what the example was from the research that Shopify provided and it'll it'll maybe explain it a little bit so what they found there was one case where they found one store they were trying to see which skin care brand or soft skin would get recommended by the AI.

21:58.247 --> 22:20.612
[SPEAKER_02]: In their research, they found that more often than not, for no other reason that they could associate to, it wasn't like this brand was ranking higher on Google or, you know, it wasn't, it wasn't didn't have particularly a nicer page or anything even like that or didn't have nicer reviews or didn't have the communication, you know, everything else is about equal.

22:21.112 --> 22:27.398
[SPEAKER_02]: It was recommending this product, which was a skincare brand by Haley Bieber, I can't remember the name.

22:27.939 --> 22:39.630
[SPEAKER_02]: And when they prompted the AI to figure out why it was recommending it, then their query was, you know, I want skincare to help with my, you know, rough skin.

22:40.090 --> 22:44.374
[SPEAKER_02]: The AI quoted to say that, well, it seems like you would like this because

22:45.035 --> 22:53.462
[SPEAKER_02]: it recorded the founder, which was Hailey Bieber, saying that I want skin so soft, you can take a bite out of it.

22:54.744 --> 23:06.174
[SPEAKER_02]: And that statement really stuck with people because that was a case, and they found other examples of this, where that statement is almost like a anti-AI, where it's like

23:07.179 --> 23:12.986
[SPEAKER_02]: That's a weird thing to say, you know, skin so soft you'd want to take a bite out of it.

23:13.346 --> 23:18.031
[SPEAKER_02]: You know, and AI would never tell you, hey, you know, I'm going to make your skin so soft.

23:18.352 --> 23:23.698
[SPEAKER_02]: You'll, you'll want people want to bite it, you know, like it just today wouldn't forget that.

23:23.938 --> 23:27.222
[SPEAKER_02]: So it looked at that and said, oh, that's interesting.

23:27.662 --> 23:29.945
[SPEAKER_02]: So this is the example of taste.

23:29.985 --> 23:34.832
[SPEAKER_02]: And not, you know, tasting the skin here, but taste taste in terms of uniqueness.

23:35.332 --> 23:40.439
[SPEAKER_02]: And that's one of the tips I would tell people is be unique and creative.

23:40.559 --> 23:43.944
[SPEAKER_02]: And it's appears the research showing, but the AI is,

23:45.530 --> 23:48.471
[SPEAKER_02]: preferring the unique things to show.

23:49.451 --> 23:52.352
[SPEAKER_02]: And my theory, and nobody really knows why.

23:52.392 --> 23:55.053
[SPEAKER_02]: This is happening, maybe other than the open AI researchers.

23:55.433 --> 24:08.877
[SPEAKER_02]: But the theory is that because these LLMs are trained on social interactions and probably being weighted in some regard to the number of likes or number of upvotes, or that when there are things that

24:10.337 --> 24:15.159
[SPEAKER_02]: interesting to people, they are kind of ranking higher.

24:15.259 --> 24:23.861
[SPEAKER_02]: So it's thinking that, hey, that's probably something you like because weird stuff like that has been uploaded in the past and that's how it's trained.

24:24.282 --> 24:25.042
[SPEAKER_02]: That's a theory.

24:25.602 --> 24:34.425
[SPEAKER_02]: Now the other theory you could say if you want to get fun is that it's because the AI is trying to understand what it means to be human and like trying to think about, you know,

24:35.205 --> 24:48.394
[SPEAKER_02]: Why would somebody want to take a bite out of the skin to so then it just decides like, okay, I'll show this man doesn't make sense to me, but I'll show it, you know, yeah, so be unique and they call it taste because it's like, you know,

24:56.700 --> 25:16.033
[SPEAKER_02]: And AI seems to prefer it, seems to give that if there's 10 other skin care brands that solve the problem, but one of them says that it's so soft that you can take a bite out of it, it says, well, if you really want soft skin use the hallowed beavers can brand otherwise, you know, maybe these are the other ones and that's what you want you want to be that one recommendation.

25:16.613 --> 25:17.293
[SPEAKER_02]: Does that make sense?

25:17.373 --> 25:21.834
[SPEAKER_00]: Like, it makes sense because it's also given us all plenty to think about.

25:21.854 --> 25:33.236
[SPEAKER_00]: So everyone, you need to get your specificity or social proof and your taste sorted to win that AI search at the moment, who knows what will come in twenty twenty seven, but it's, I think this is a pretty good bet for the future.

25:33.717 --> 25:35.897
[SPEAKER_00]: Jay, thank you so much for taking us through all of that.

25:36.017 --> 25:40.158
[SPEAKER_00]: Listeners, make sure you stick around for the top tips round for lots more insights.

25:41.980 --> 25:46.982
[SPEAKER_00]: Okay, Jay, so far we've gone deep into how to get AI to recommend your products.

25:47.062 --> 25:50.523
[SPEAKER_00]: Now you get 12 hours for your inside and knowledge about the whole of AI search.

25:50.863 --> 25:56.965
[SPEAKER_00]: So for the following questions, you aren't going to be anything to do with AI search, which of course includes everything we've already been talking about.

25:57.025 --> 25:58.906
[SPEAKER_00]: Jay, are you ready for the insider tips?

25:59.846 --> 26:00.767
[SPEAKER_02]: Yes, ma'am.

26:01.247 --> 26:03.288
[SPEAKER_00]: Okay, let's start with Neuby advice.

26:03.708 --> 26:09.370
[SPEAKER_00]: If we even inspired someone to take their first step, what do they need to know to give themselves the best chance of success?

26:10.418 --> 26:17.883
[SPEAKER_02]: think about creating a really cool brand and what your brand and your customers and speak to your customers and focus there.

26:17.963 --> 26:20.866
[SPEAKER_02]: Those are the steps, don't try to optimize for it.

26:20.946 --> 26:23.808
[SPEAKER_02]: I add this stage, which is a weird thing, right?

26:23.828 --> 26:25.109
[SPEAKER_02]: So, you can't even do this.

26:25.609 --> 26:30.713
[SPEAKER_00]: Once you've started, of course, though you've got to keep optimizing so what's your favorite way to improve performance?

26:32.292 --> 26:54.916
[SPEAKER_02]: favorite way to prove performance on the AI search would be to, you know, I don't know if this is the best way, but my favorite way of, you know, from an entrepreneur stands would be to get fun, like the taste portion where I said one of the three things they said is making differences taste, which is kind of just the fun part of entrepreneurship in my opinion.

26:54.936 --> 26:55.816
[SPEAKER_02]: You get to be creative.

26:56.116 --> 26:56.956
[SPEAKER_02]: Be more creative.

26:57.356 --> 26:59.757
[SPEAKER_02]: And I really appreciate by the way that you've

27:01.177 --> 27:20.257
[SPEAKER_02]: been through this where it commerce ended up becoming like a whole sea arrow game and everybody was just optimizing for the little thing and I hated that I felt like it took all the creativity out but it made sense you know I was like it made sense dollars and sense I got it you know but it really took out all the creativity and I'm one thing I'm happy about is that

27:20.797 --> 27:22.818
[SPEAKER_02]: you know, maybe that's coming back to making it come back.

27:22.838 --> 27:25.140
[SPEAKER_02]: We can be creative about our accomplished experiences.

27:25.580 --> 27:27.781
[SPEAKER_02]: So that's what I would say.

27:27.881 --> 27:34.665
[SPEAKER_02]: My answer to that is you got to be creative to wow up and it optimizes creativity, optimizes, which is a great thing.

27:35.025 --> 27:41.429
[SPEAKER_00]: And I can't help but think as well as being the fun, but it's probably the bit that's got the longest legs, the differentiating factor.

27:41.869 --> 27:45.211
[SPEAKER_00]: But if someone if someone listening to this wants to learn

27:48.964 --> 27:58.552
[SPEAKER_02]: You know, as I said earlier on, everybody is trying to grip their own product, but Shopify has a very unique position in that they're not trying to gain, not aggressively trying to gain more market share.

27:58.852 --> 28:04.397
[SPEAKER_02]: So their research on a Gentic is a little bit better than most people.

28:04.417 --> 28:11.743
[SPEAKER_02]: They have the most to lose if a Gentic search, if their sites are not optimized, a Gentic search, and everybody's already using them anyway.

28:11.783 --> 28:12.443
[SPEAKER_02]: So they don't need to

28:13.264 --> 28:16.145
[SPEAKER_02]: They're not only to sell you on their authentic features, they just got to do it.

28:16.225 --> 28:29.110
[SPEAKER_02]: So I would follow the Shopify blog as a free resource and pay attention to their YouTube series and they'll provide unbiased data bits on what you should actually be doing without selling you something.

28:29.613 --> 28:30.094
[SPEAKER_00]: Yeah.

28:30.274 --> 28:31.115
[SPEAKER_00]: I totally agree.

28:31.155 --> 28:37.485
[SPEAKER_00]: I think that's a great place to keep an eye on the reality of it with a fairly unbiased source, like you say.

28:37.886 --> 28:44.657
[SPEAKER_00]: And finally, dare I ask Jay, crystal ball time, anything we need to be getting ready for in the next six to 12 months.

28:45.692 --> 28:46.312
[SPEAKER_02]: Yeah.

28:46.412 --> 29:12.518
[SPEAKER_02]: I think images product photos are something that we're just going to see a tsunami of confusion and change is that we're noticing already in and our data is that more and more images are a generator or an AI editor and governments are trying to get, I think I think the EU started to put in rules where you have to label, label images, but what happens when every image is labeled

29:14.558 --> 29:29.946
[SPEAKER_02]: It's going to be, I believe, is going to be the train, you know, the train is moving towards every image being AI generated in some regard, the I edited and I think that brands need to start thinking about how do I differentiate out of that.

29:29.966 --> 29:36.130
[SPEAKER_02]: I think we might see a comeback for modeling as my theory, you know, like, especially when the model is unique to your.

29:37.150 --> 29:43.060
[SPEAKER_02]: customer, or when you can, the model relates the best with your customers, but I think brands need to start planning for that.

29:43.120 --> 29:47.106
[SPEAKER_02]: That every image, everything you see online, you're going to assume it's fake.

29:47.747 --> 29:50.892
[SPEAKER_02]: And how do you present to the customer something that they know isn't fake?

29:51.670 --> 29:54.232
[SPEAKER_02]: you know, when they assume regardless of the label.

29:54.772 --> 29:56.714
[SPEAKER_02]: So yeah, that's my crystal ball.

29:56.754 --> 29:59.496
[SPEAKER_02]: That's six to 12 months is that's going to be a big thing.

29:59.796 --> 30:01.998
[SPEAKER_00]: That is going to be a big thing and I haven't.

30:02.078 --> 30:04.459
[SPEAKER_00]: Yeah, we should wait and see to see that one guys.

30:04.940 --> 30:06.481
[SPEAKER_00]: Jay, we are nearly at the end of the show.

30:06.881 --> 30:15.768
[SPEAKER_00]: So could you please let the listeners know where they can find you and a cello if they want to getting in contact with you and get help with their marketplace listings?

30:16.763 --> 30:20.244
[SPEAKER_02]: Yeah, so you can go to our website, ocelo.com.

30:20.804 --> 30:34.449
[SPEAKER_02]: If you are looking to sell on Nordstrom, Macy's Target Plus or Blooming Dales, Marketplaces online, or if you're already selling on them, and you're struggling with listing, you're having to fill out spreadsheets, make,

30:35.209 --> 30:54.927
[SPEAKER_02]: little image edits here and there and keeping yourself up at night, you know, with that and you're getting delays, especially as we enter, you know, especially into the holiday season, then please do check us out, take care of your fashion brand and, or you can find me on LinkedIn, I'm always posting there and your social spots.

30:55.932 --> 31:01.654
[SPEAKER_00]: So, Ocelo.com, if you want to find out more about it or go and hunt down Jay and start stalking him on LinkedIn.

31:02.334 --> 31:11.677
[SPEAKER_00]: For all the latest insights, Jay, thank you so much for coming on the podcast, as always, I have learnt a ton from you and I'm really glad our listeners have got you too, so thank you for being here.

31:12.358 --> 31:13.198
[SPEAKER_02]: Thank you for having me.

31:15.575 --> 31:24.442
[SPEAKER_00]: Well, plenty there for us all to get a stuck in teeth stuck into, especially trying to work out what does taste mean for your brand and for your products.

31:24.903 --> 31:26.924
[SPEAKER_00]: And great to get to catch up with Jay about all that.

31:26.944 --> 31:28.145
[SPEAKER_00]: I said, he's fascinating guy.

31:28.165 --> 31:29.807
[SPEAKER_00]: I always love it when we get time to catch up.

31:30.127 --> 31:32.129
[SPEAKER_00]: And especially when we get to share it with all of you.

31:32.569 --> 31:35.551
[SPEAKER_00]: So what you need to do now is you need to go and get your heads around.

31:36.352 --> 31:38.293
[SPEAKER_00]: find the all important piece of paper.

31:39.173 --> 31:57.582
[SPEAKER_00]: Specificity, the social proof and the taste elements, how can you bring those onto your product pages, bring those into your marketing activity, leverage your customers and so forth to create them in order to be more likely to be selected by the AI as the right choice for the person doing the searching.

31:58.322 --> 32:04.189
[SPEAKER_00]: You get the links to all those things we discussed, the full transcript of the episode important notes more at keepoptimizing.com.

32:04.590 --> 32:12.819
[SPEAKER_00]: You can also use a special direct episode short link that's keepopt.com forward slash episode number to get to the right page on the site.

32:12.839 --> 32:15.562
[SPEAKER_00]: When you get there, please do add yourself to our email list.

32:16.163 --> 32:19.805
[SPEAKER_00]: Thank you for tuning in to this episode of the Keep Optimizing podcast.

32:20.125 --> 32:23.947
[SPEAKER_00]: You've enjoyed this one and want to continue learning about AI search right now.

32:24.287 --> 32:25.507
[SPEAKER_00]: Then pick up our thread.

32:26.028 --> 32:29.929
[SPEAKER_00]: Scroll back up your feed to episodes 294 to 298 from earlier this year when we did a series on both SEO and GEO.

32:35.812 --> 32:42.780
[SPEAKER_00]: including what to do on your category pages and a deep dive into product data and much more, well worth the list in those five episodes.

32:43.281 --> 32:52.772
[SPEAKER_00]: You can also find our full thread of AI search content via the shortlinkkeepop.com forward slash SEO because yes, we're keeping it with the SEO stuff for now.

32:53.793 --> 33:01.235
[SPEAKER_00]: And please, do tell your fellow marketers about the show because I want to help as many of you as possible to improve the performance of your e-commerce marketing.

33:01.715 --> 33:07.136
[SPEAKER_00]: Have a great week and make sure you listen to the next episode so I can help you to keep optimising your marketing.

