#648 – Is AI Ruining Amazon?

Is artificial intelligence reshaping Amazon selling for the better or leading it astray? In this episode, we scrutinize the nuanced impact of AI on Amazon sellers, posing the provocative question of whether AI is indeed ruining Amazon. We dive into the labyrinth of AI’s potential benefits—boosting sales and streamlining operations—while also exposing the risks of over-reliance on AI without human input and data. As AI tools like ChatGPT become more prevalent in product and keyword searches, we emphasize the crucial role of specialized tools such as Helium 10 in ensuring accuracy and effectiveness.

Our discussion highlights the pitfalls and promises of using AI for Amazon keyword and product research. Through real-world examples, we dissect the challenges that arise when relying solely on AI-generated data, underscoring the necessity of feeding AI robust and relevant information. Exploring the risks of AI-written Amazon listings, we stress the importance of informed technology use to achieve optimal results. By illustrating the limits of AI tools in identifying trending products and optimizing listings, we advocate for a balanced approach—melding AI’s capabilities with human oversight and specialized data.

Finally, we turn our focus to enhancing the synergy between AI and Helium 10 for your Amazon brand’s success. Through a case study of a wooden snack organizer, we demonstrate the power of integrating comprehensive data with AI to boost keyword indexing. With insights from Andrew Bell’s customized GPT tool, we reveal pathways to even greater performance. We invite listeners to engage in this conversation, sharing their thoughts and experiences on enhancing AI integration to better serve the community’s needs. Whether you’re a seasoned seller or a curious newbie, this episode offers valuable insights into harnessing AI’s power with precision and care.

In episode 648 of the Serious Sellers Podcast, Bradley talks about:

• 00:00 – AI Usage in Amazon Selling
• 03:44 – AI’s Impact on Amazon Sellers
• 06:56 – Using Amazon Tools for Product Research
• 12:08 – High Demand Snack Bar Organizer Product
• 15:09 – Amazon Algorithm Changes and Keyword Indexing
• 18:50 – Importance of Keywords in Listing
• 19:09 – Optimizing Amazon Listings With AI
• 21:57 – Keyword Optimization for Amazon Listings
• 27:29 – Optimizing Amazon Listings With AI + Helium 10 Data
• 31:22 – Andrew Bell’s GPT Boosts Listing Performance
• 35:18 – Enhancing AI Integration for Brand Success

Transcript

Bradley Sutton:

Is AI ruining Amazon? Today, we’re going to talk about this topic and talk about the ways that you should be using AI, some of the ways that you shouldn’t be using AI, and how using it in the right way could even increase your sales by 30% or more. How cool is that? Pretty cool, I think. Hello everybody, and welcome to another episode of the Serious Sellers Podcast by Helium 10. I’m your host, Bradley Sutton, and this is the show that’s completely BS-free, unscripted and unrehearsed organic conversation about serious strategies for serious sellers of any level in the e-commerce world, and today we’re going to talk AI. Now, we did probably like a click-baity title to this, like is AI ruining Amazon in a TLDR way? To answer that. The answer is no. But what are we talking about here? All right, because there actually is a circumstance. I feel that using AI in the wrong way is, definitely could and is harming a lot of Amazon sellers and actually costing you a lot of sales. Now, what do I mean when I say is, you know, AI ruining Amazon? Now, there’s a lot of ways that you can take that. Let me talk about what it’s not. Now, I’m not talking about what we’ve talked about before in the weekly buzz how sometimes AI on Amazon is actually changing listing titles and all of a sudden there’ll be error messages that says I’m sorry, I cannot fulfill this request because the Amazon AI is trying to actually redo a certain title or something like that. That’s not what I’m saying. I’m not also today talking about Rufus. All right, Rufus is like this AI on the buyer side of Amazon that a lot for a lot of people Most people kind of sucks like, doesn’t give the right answers, hallucinates a lot, not very helpful. That’s also not what we are talking about today. I’m talking about as Amazon sellers using AI. How can that potentially ruin Amazon for you and is AI bad? Right, and I’ll just let you know right off the bat I use AI. I wouldn’t say a lot, but I use AI a significant amount, whether we’re talking about for the podcast, for my Amazon businesses, for Helium 10, et cetera. As a Helium 10 employee, I’m never going to say AI is bad, why Helium 10 actually has been using like machine learning that today you know some aspects of it can be considered as what some people call AI. We’ve been using that since like 2018, 2019.

Bradley Sutton:

Inside of Helium 10, there’s multiple ways to use AI, like, for example, do you know that if you hit the magnifying glass on the top right of the screen in Helium 10, it pulls up a whole AI kind of interface and I can ask questions like how can I use Cerebro to find the best keywords for an agent? AI is going to go ahead and give a very detailed answer. Multiple tools in Helium 10 use AI. In Cerebro there’s an AI filter where I can write stuff like hey, show me only the Spanish keywords that are showing in the search results and here for this product I can see a whole bunch contenedores para snacks, et cetera, et cetera. So now all of a sudden, in an instant, it filtered out all of the Spanish keywords. We’ve got tools like product launchpad, where you organize all of your product research and then, based on all the data points, AI is going to let you know like a summary of what’s going on. We have, even at on our base plan, the platinum plan, access to tools like AI advertising, where it’s like a hands-off way of managing your advertising, your PPC. You just put in, like your ACoS goal and what product, and the AI like manages everything for you. So obviously you know AI is a good thing if used correctly. Helium 10 is using it. We’ll use it even more in the future. So, no, AI is not ruining Amazon just in that strict sense. But what I’ve been seeing lately which is why I’m doing this kind of podcast last second here is I’m kind of scared about some Amazon sellers making the wrong decisions out there because they’re using it in the wrong way and they’re not understanding the limitations of AI.

Bradley Sutton:

Um, I just gave you some Helium 10 examples. There’s a billion examples out there, not a billion, hundreds of thousands of different ways that you can probably use AI for your Amazon business. That has nothing to do with Helium 10. You know, like image generation. By the way, obviously, Helium 10 has image generation with AI. There’s way more stuff, even that Helium 10 does with AI that I didn’t even mention. There’s ways that you can write email flows for your customers for your Shopify website. You can actually create websites out there with AI. You can have it manage your customer service. There’s a lot of great ways. But what’s happening is I see some customers getting what I almost want to say lazy right and they’re relying on AI too much without the right context, and this is where it can be dangerous. So, for example, I’ve seen people say, oh yeah, I don’t need Search Query Performance in Amazon, I don’t need product opportunity explorer, I don’t need Helium 10 in order to do keyword research, I’m fine with just ChatGPT. And let me show you why that can be dangerous. Like, for example, I put in a search here to ChatGPT. This is an unfiltered search. This is literally what you can see exactly on the screen. What I typed in, I said what are the five highest search volume search terms that could bring a product that is, a bat shaped bath mat sales on Amazon? And, spoiler alert, I actually looked in a Search Query Performance. By the way, here you’re getting a sneak peek at Helium 10’s Search Query Performance Analyzer tool. I can see all of the keywords that brought our bat bath mat sales and the top ones are bat bath mat right. Next one Batman bathroom rug. Next one bat rug. Spooky bathroom decor, bat bathroom decor. These are like the top five keywords. Uh, there’s some more here. Bathroom set bats. Um, gothic bathroom mats. There’s like 10 keywords that brought me sales in the last couple of weeks. Um, again, shout out to the team who put together this Search Query Performance tool. Uh, you guys who are diamond members will probably see this in your account next week. Anyways, that’s what the answer should be right.

Bradley Sutton:

But now let’s look what ChatGPT says. Um, it says, hey, to effectively market a bat shaped bath mat on Amazon, targeting high search volume keywords is crucial. Blah, blah, blah, blah. What are the ones that it says could bring a product bat shaped bath mat sales? The top Gothic bathroom decor that’s actually not bad. Uh, that hasn’t brought me sales in a while, but it technically could. Halloween bath mat Actually not bad. That hasn’t brought me sales in a while, but it technically could. Halloween bath mat. Okay, Bath bathroom rug, Novelty bath mat. Very, very generic keyword there. Black bath mat Even more generic keyword. So keyword research just with ChatGPT and not feeding it the data that it needs is not going to set you up for success. Now here’s funny. I swear guys, you can see this. I am not like a Photoshopping this. This is my actual screen. This is how it concluded. I didn’t give it any other prompts, but after it gave me those five keywords, what did ChatGPT say?

Bradley Sutton:

To identify the most effective keywords for your specific product, consider using Amazon keyword research tools like Helium 10 Cerebro, which can provide insights into search volumes and keyword relevancy. Oh, my goodness, we need to make ChatGPT a Helium 10 affiliate here. It actually is giving us a shout out for free, but again, this is actually the point. It’s like, without putting extra data like actual search volume numbers, keyword sales estimates, historical rankings, sponsored rank history, ChatGPT is going to be limited. It doesn’t have like some universal access to like Helium 10’s database or to Amazon’s backend numbers and API and stuff, so it’s not going to be able to give you, like you know, the most accurate information that’s actual, it’s not going to be wrong all the time. You know, because of all the data it has from like Google and other sources, you know it’ll give you some good keywords, but again, without putting the right input, it’s kind of dangerous to rely on this for some aspects of Amazon. And even ChatGPT itself says, yeah, you probably need Helium 10 out there. That probably should have said hey, you need Amazon Search  Query Performance opportunity to explore, et cetera.

Bradley Sutton:

Another aspect, um, that you got to be careful of is in product research. All right, there’s a lot of publicly available data on amazon.com, right, like, it shows the products. You can see the listings, you can see number of reviews, a lot of stuff. I don’t know how much access ChatGPT and other um. You know AIs have to this, but you know it know it might be substantial, right, but you can’t just rely on ChatGPT or any of these AI things to just say, hey, find me a new product you know to sell on Amazon. That’s low competition because it’s not going to have all of the critical data points that you might need.

Bradley Sutton:

Now I know people who have, like, done some really cool prompt and actually come up on some trending things that weren’t even on Amazon yet, but it found it off of Amazon. Hey guys, guess what that’s? That’s how we launched the whole project exiting with coffin shelf. It wasn’t even on Amazon. We found it off of Amazon, like on Etsy and Pinterest. So, yes, can you, without using any tool or even Amazon’s opportunity explorer tools or Helium 10, potentially find a product to sell on Amazon? Yes, but that’s not necessarily the best way. Like, for example, here is a prompt I did on ChatGPT. It says, or I said I’m looking for a trending product in the Amazon kitchen and dining category that I can launch. It needs to be in a niche where some of the top sellers have less than 150 reviews.

Bradley Sutton:

I wasn’t sure if it could do this, but I was like let’s try it when the main keywords of the product are easy to rank for and where the top product size here is at most large standard size. And it gave a few things here. But just by looking at it I was like I don’t think this is really anything that exciting. Here We’ve got high quality baking sheets with grids versus how ceramic pasta bowls, you know? You know we have to have search volume behind these. Let’s just take a look at the one I put here high-quality baking sheets with grids. Like I doubt high quality is even a search term and sure enough it had zero search volume, right? So, if I just search baking sheets with grids on Amazon, let’s take a look at the search volume there. Also, zero searches. So, I don’t know what’s going on like this. This is not a great, you know, product opportunity here, where most of these products here on page one have thousands of reviews, all right, so thousands. That already didn’t do what I told it to, and this is not even a high search one. But again, you can’t really blame ChatGPT. It’s not like a flawed. No like, how could it possibly know what’s high competition? It doesn’t know all of the review counts. It doesn’t know you know what’s easier to rank for. But again, if you fed it the right data, now all of a sudden ChatGPT becomes a powerful weapon. So just be careful of that.

Bradley Sutton:

Like I actually did a search in Black Box for something very similar in Helium 10 Black Box. I’m like hey, show me something between 2000 and 7,000 searches. It’s got at least three words. The price is between 20 and $80. Uh, at least seven out of the first 10 products have less than 150 reviews. It’s in the kitchen and dining category. It’s got a title density of four, meaning only four listings on page one have that exact search term. And I found a lot of good keywords. And actually, for the rest of this episode I am going to be diving deep into this snack bar organizer. All right, that was something I had never heard of and that Black Box found for me. And for the rest of this episode I’m going to talk about the most important aspect of how people are misusing ChatGPT and misusing AI, and that is with writing listings. All right, this is like the biggest danger, in my opinion of what I see happening out there, and probably the most common thing that people are using AI for when it comes to Amazon is making listings, and I’m going to illustrate just how dangerous it is. If you’re using AI in the wrong way. Want to enter in an Amazon keyword and then, within seconds, get up to thousands of potentially related keywords that you could research, then you need Magnet by Helium 10. For more information, go to h10.me. forward slash Magnet. Magnet works in most Amazon marketplaces, including USA, Mexico, Australia, Germany, UK, India and much more.

Bradley Sutton:

So take a look here those of you watching this on YouTube a snack bar organizer what kind of products it is. It’s actually a pretty, pretty interesting looking niche. I’m not sure if it’s too late, but you look at the brand analytics top clicked and top converted. One is this product that’s hasn’t been around too long? It only has barely over 100 reviews it’s been doing. Let’s just take a look at the sales here $40,000 in the last 30 days, all right. And the second product that comes up on this page, Brand Analytics number two, only has four reviews. The Brand Analytics, number three on this page only 70 reviews. So Black Box really found a good product here. That is pretty new. It’s the snack like bar, um, like organizers where you can put like chips and stuff trying to describe it for those of you just listening in your car or something, but anyways, like right away. Um, on the second line of search results, I’m already seeing a lot of irrelevant products. Uh, that aren’t even what this is. So this is a pretty decent thing. But let’s go ahead and make a scenario of hey, we want to sell this product and we want to make a listing. All right, so I did this case that you took me all night tonight to do this, where I did different scenarios.

Bradley Sutton:

Now, the first thing is to have an equal playing field is let’s go ahead and use Helium 10 to find all the relevant keywords. So some of the steps I took was I use Cerebro and I put all the top competitors that were super relevant and then I started extracting keywords. Like I hit the top keywords button, where are most of the products ranking for getting their sales from? Where are some of the uh, the keywords that only a couple of the products are ranking for? Where are the ones where everybody’s is ranking high in sponsored results, meaning they’re all bidding on top of search. Then I threw it into the Helium 10 Black Box tool. It’s like hey, show me all of the keywords, guys. If you don’t know how to do this, you’re leaving money on the table, you Helium 10 users. I said show me all of the keywords that brought guaranteed sales direct from Amazon, to these products, the top competitors. Like this is not a Helium 10 estimation, this is directly from Amazon Brand Analytics. And there was actually 44 keywords that, um, amazon is saying that brought sales to these products in the last 15 weeks. There’s obviously more than that, but there’s 44 of them where these products were one of the top three clicked and they had sales.

Bradley Sutton:

So I created this whole keyword bank or keyword list, and there was like let me actually just pull it up here and there was actually this total list of keywords. I think it was about 75 keywords that came out, came out and some of them are like stuff that is, you know you wouldn’t maybe expect, like, for example, cool things for movie room, snack storage box, uh, snack bar, treat yourself, organizer for fans of parks and rec right there, movie theater, snack bar station, et cetera, et cetera. So these were all the top keywords. Probably only took me like 10 minutes to do it. I could have done some more extensive keyword research, but I just want to have like a baseline, all right. So in 10 minutes, what are the keywords that I could find that are relevant to this niche that could bring sales? All right, now for some level setting here. This is a very important point.

Bradley Sutton:

Amazon search has always changed for 10 years. The algorithm changes. How you get indexed sometimes changes. Different parts of listings have different impacts. It always changes, right, but the part that’s probably not going to change anytime soon and that’s always equal, is that in order to have the best chance to be indexed for a keyword, you need to have the keyword in your listing. Now, as different AIs and algorithms kind of like advance on Amazon, you’re going to see more and more. I predict and it’s not really me going out on a limb this has actually been happening for years already where sometimes after a while, amazon learns the behavior, learns more about what a product is.

Bradley Sutton:

You’ll see products that are indexed for keywords that are not even in the listing. You know that can happen because of how it performs in advertising and PPC, like Amazon shows you an auto campaign for a keyword you don’t even have in your listing and then there’s good interaction with customers. You’ll get indexed for that keyword. Or Amazon sees a word in a review, or maybe they look at an image and there’s something you don’t have anywhere else in your listing but they can kind of extrapolate that this product is about it. Yes, you could potentially get indexed, but still, the number one way to get indexed is to have the keyword in the listing.

Bradley Sutton:

The number one way to get indexed is to have the keyword in the listing. Like, for example, here’s some of the top keywords for this snack bar niche. And what I did was I look at the very top product, the product that is the number one Amazon’s choice and the one that, that that is selling 20 units a day. And if you take, uh, using Frankenstein, you check for the individual keywords, uh, like snack, all right. And then, if I examine the all the texts that’s on the listing, all right, whether it’s from the backend search terms, whether it’s in the title bullet points, I have this kind of tool I use where I could see all of the texts in a certain page. Right, if I start searching for these, you can see snack is 28 times in this listing. If I start searching for these, you can see snack is 28 times in this listing. Let’s see Storage is 10 times in this listing. Holder, which is one of the top part of one of the top keywords. It’s all over the listing. So, again, the best way to get indexed and be searchable and then later rank for it is you have to have that keyword in your listing. All right, that’s what’s going to give you the best chance from day one to start ranking for keywords. All right, uh, that’s what’s going to give you the best chance from day one to start ranking for keywords.

Bradley Sutton:

Conversely, another thing I did was I went into Cerebro and I was like let me take one listing and see the keywords that it’s not ranking for but other competitors are getting sales from it, and which of those keywords maybe are not indexed. So I have a whole bunch of keyword phrases here that one product is not indexed for. Obviously, it’s not ranking for it. If it’s not indexed, it’s not getting sales for it, like, for example, black carousel, rotating snack tray, breakfast station organizer. These are all keywords that are bringing sales to this other snack organizer, but this one product is not, and so what I did was all right. What is going on like why didn’t, why aren’t, why isn’t it indexing? All right. So I take some of those individual keywords. Let’s take carousel, for example. Let’s go back to the page that analyzes everything that is on the page. Carousel. Do you see, I don’t know if you guys can see it here it came up zero times. The word carousel is not in the listing. Conference is not in the listing and it’s not index Theater let’s go ahead and look for that. Theater is not anywhere in the listing. So you guys see the picture here.

Bradley Sutton:

The moral of the story is rule of thumb If you want to be indexed for a keyword and searchable, you have to have the keyword in your listing to start. That’s what gives you the best chance. Is that 100% necessary? No, gives you the best chance. Is that a hundred percent necessary? No, down the road you could get indexed for other keywords. It’s not. But from day one it’s very unlikely. You’re going to be indexed for a keyword if you don’t have in your listing. Uh, and conversely, if you’re expecting to get sales, you probably already had that keyword in your listing. All right, so? So I showed you the proof of that. All right, so? So, now that we understand that, what happens when you use AI to make a listing. Because it doesn’t take much. Like let’s use the first scenario.

Bradley Sutton:

All right, what I did, I’m going to show you exactly. I still have the window open right here. I just did this a few minutes ago. I downloaded some pictures of one of the snack organizers, okay, I uploaded it to ChatGPT and this is exactly what I said. I was like hey, I want to make a well-optimized, SEO said I was like hey, I want to make a well-optimized, SEO friendly Amazon listing for a new product I’m going to sell. It’s a wooden snack organizer for any countertop. Uh, I attached a couple of pictures of what the product looks like. Uh, looks like, and I even copied some texts from the listing itself to like hope that it fills out some of this, this data. I says hey, it’s got four compartments and a drawer. The front three compartments can be used to store snack bars, candy, and the backspace is large enough to store drinks and disposable cups. Please write an Amazon listing for me, including the title, bullet points and description. Per best practices for Amazon listings, all right, so I did a fairly. This is not the perfect prompt here, but this is a pretty, pretty detailed one.

And sure enough, you know, ChatGPT came up with a listing that looked not bad. All right, like the title. It came up with snack bar organizer for countertops. Wooden snack storage with four compartments and pull-out drawer. Large capacity snack caddy for home office, break room and coffee bar holds snacks, drinks, cups and more, and it gave bullet points, it even had emojis, it gave a full description, et cetera. Okay, and it was like, hey, we used relevant keywords like snack organizer to maximize visibility on Amazon. Let me know if you have any tweaks. All right, so that was what ChatGPT said.

Bradley Sutton:

So now the next step is I went into Listing Builder in Helium 10. That’s why I’m wearing my LB Listing Builder hat. And what I did was I imported all of the competitors, the top competitors in the niche, so I can get like a baseline score and see who are score and see what are the main keywords and everything. And then I imported those 75 keyword phrases and then I just copied and pasted the ChatGPT listing into Listing Builder. What ChatGPT came out with and what does this do? This helps me, first of all, see how optimized for SEO it is based on these keywords and this score. The number really doesn’t matter. This score just happens to be 264,000, but it was ranked seventh out of eight. All right, so out of the eight or the seven competitors that are out there, plus mine or this ChatGPT, listing this one ranked next to last. Now, that’s not saying too much. The score is just kind of like, meant to show how many of the keywords you used, uh, used, and what kind of search volume they had. But the main key here, guys, is the number of keywords I’m going to be indexed for and thus able to rank.

Bradley Sutton:

There were 75 phrases, if I sort it by search volume, out of the top 10 highest search terms that are bringing sales to the competitors in phrase form, ChatGPT only had three of them in phrase form. In phrase form, ChatGPT only had three of them in phrase form. If I sort this by the top 10 most relevant keywords, regardless of search volume, ChatGPT only had two out of the top 10 most relevant in phrase form. Now here’s the kicker out of all those 75 phrases, there’s actually only 85 individual keywords in this list. Remember 75 phrases is the keywords that are driving sales to the existing competitors or there are signs that it’s highly relevant. Those probably have like 200 words. Like, maybe each of those keyword phrases might have two words or three words they’re not like one-word phrases, right, but a lot of them have the same exact phrase, like, for example, we see snack bar organizer, we see snack bar station, snack organizer. You see how there’s the same words coming up. So if you want to count the individual words this is another thing listing builder does there’s actually only 85 individual unique words.

Bradley Sutton:

So theoretically, to be searchable for all of those 75 main keywords, I only need to have 85 individual words, even though there’s like 200 and something words in there right Now. Here’s the thing ChatGPT only indexed me for one, two, three, four, five, 21 out of those 85. That means for these 60 other keywords that are making up a lot of those phrases, the fact that it’s not there, that means that more than half of the phrases that I need to index for to get sales, this listing it’s probably a great looking listing. It did a pretty good job. Like kind of like optimized somewhat for buyers, right, but for Amazon search guys, I’m leaving more than 50% of my sales on the table because it is not even helping me be searchable for more than 50% of the relevant phrases and 60, 75% of the individual keywords that I need to index, for I am not going to be indexed. For you see the danger here, guys, by just relying on ChatGPT without giving it any input at all, right? So hopefully that’s, that’s pretty clear to you guys.

Bradley Sutton:

Now there’s actually some ChatGPT whizzes gurus, if you were out there, who make some Amazon listing creation ChatGPTs that are, like, designed to work more for Amazon. But again, there is a right way to use these and a wrong way. Um, one of the hands down best um ChatGPT creator out there for Amazon sellers is a guy named Andrew Bell. You just type in Andrew Bell ChatGPT, you’ll see hundreds of his GPTs that he has handmade for Amazon sellers. All right, so here I went and I went to his ChatGPT channel and I’m using his Amazon optimization GPT and with it I went into his GP GPT and it says describe your product to me. And so what I did was I did the exact same thing that I did in the regular chat GPT that didn’t have any fancy filters. I uploaded the same two images. I put the same exact copy or the prompt for it. You know about, hey, I need an SEO-friendly Amazon listing. And then Andrew’s GPT gave me some output and again it’s not bad. You know, it’s giving me a pretty good listing GPT here, but let’s look how it did in Listing Builder, all right. So I put that Andrew’s GPT optimization all in here.

Bradley Sutton:

Uh, the score was actually better, I think, than the regular GPT, or actually. No, it was about the same or a little bit worse. Actually 256,000. Again, the number is not that important, it was all it was. It ranked dead last, though instead of seven out of eight, but again, that’s not important. Seventh or eighth is not important. The important thing is how many keywords did it index me for on its own. Out of the top 10 search volume? It again indexed me for three in phrase form Out of the top 10 that were most relevant competitor performance score. It had, in phrase form, two of those top 10. So very similar to ChatGPT and it did a little bit better. I think. It indexed me for about 25 keywords out of to ChatGPT and it did a little bit better, um, I think. I think it indexed me for about 25 keywords out of the 85 here, so it did a little bit better, but still, again, this to me is the wrong way to use ChatGPT. Even if I’m using an optimized like, there’s absolutely no doubt the listing is better than what the regular ChatGPT because he like spices up the prompts so that it really is a lot more optimized for Amazon, it’s a lot more optimized for buyers. But again, at the end of the day, it doesn’t matter how amazingly optimized and what kind of psychology this listing is if I’m not even going to be searchable for more than half of the keywords. So again, in my opinion, this is the wrong way to. I’m not saying don’t use this chat to be. I’m saying the way I did it, which is I didn’t feed any outside information. I just said, hey, write me a listing. I put all my relevant details there. That’s the wrong way, guys, to use AI for listings.

Bradley Sutton:

All right. Another way of using AI for listings is using Listing Builder itself, all right. So here in Listing Builder I did the same thing, had all my keywords, had the same competitors, and now I used Helium 10’s AI, which also uses ChatGPT, I believe, and I put in some of those phrases the same ones that I did with the other ones, and I put the product name and then had the ChatGPT inside of Helium 10 write the listing. Okay, now the difference here is, automatically it is using as many of the keywords that I had entered in listing builder as it could now score wise. It wasn’t that much better than the other ones. I want to phrase again or express again for the fourth time. The score is not the end, all be all of it, but yes, the score was 358,000 on this, which was good enough for fourth out of eighth. Okay, out those. So it’s better than just using raw ChatGPT.

Bradley Sutton:

But here’s the key out of the top 10 search volume phrases, how many did Helium 10 AI get in phrase form? 10 out of 10 out of the top 10, most, uh, most relevant, the highest competitor performance score one, two, three, four, five out of the top 10. It had in phrase for member. The other ChatGPT by itself only did two out of 10 for both of those, and this is the key out of the top 10 it had in phrase four. Remember, the other ChatGPT by itself only did two out of 10 for both of those and this is the key. Out of the 85 individual keywords that need to be indexed, Helium 10 was able to have 72 out of the 85 individual keywords were indexed. Only 13 was missing. Do you remember how, on ChatGPT and the other GPT, because it didn’t have those keywords, uh, it only. It was missing 60 out of 85. So for 80% of the keywords, I am now going to be indexed for just by using this, but again, this is not necessarily the best way to do.

AI listing also is just using listing builder and hitting these AI buttons without doing anything. All right, so what is the next step? Let’s say I don’t have, or let’s say I have Helium 10, but I still insist on using ChatGPT. So what I did was now I went back to that original ChatGPT, the one where it was just ChatGPT, raw ChatGPT, and in the same take a look here in the same thread. Like it just gave me the listing. Now I say, hey, this listing does not have enough of the relevant keywords.

Bradley Sutton:

Attached is a list of the top keywords in this niche. In order to be indexed on Amazon, I need to have in my listing as many of the individual words that make up the phrases in this list of keywords as possible. Can you rewrite the listing, trying to use more of the keywords on the list? And I uploaded the Helium 10, Cerebro and Black Box keywords, that list of 85 keywords that are highly relevant and driving sales to the competitors. I put this prompt and I put the description, or the description, I put the keywords, the Excel file of keywords, and it says okay, I’ll analyze these keywords in your CSV file and I’ll try and include as many as impossible. And it went ahead and rewrote the listing.

Bradley Sutton:

So what did I do? I went and I re-put the listing now into a new listing builder using all the same parameters, rewrote the listing according to ChatGPT’s new output that was based on the helium 10 keywords and the score it now is up to 300,000. All right, so remember it was only like 250,000, But here’s the key out of the top sense, uh, search volume keywords, it now had more than half of them, instead of only two. Out of the most relevant keywords, it now had three out of 10. And then, um, unfortunately though, out of the top keywords here, or out of all the individual keywords, it now had about 40 out of the 85. So a little bit better. Not 21, but about 40 now. So, um, this is, this is better. And why was it better is because I gave it the keywords.

Bradley Sutton:

Now that’s still not the end of it. This is not good. It’s still not great. I’m leaving half of the keywords on the table that I can’t get sales for until, hopefully, amazon’s own AI will index me even though I don’t have it. So I would need to go back and try and force the ChatGPT to use more keywords, right. But again, this is much better than just using GPT. Again without the Helium 10 keywords. GPT is working off of limited info. Now watch this. Now I went back to Andrew Bell’s GPT and the same thing. In the same like conversation I was having with ChatGPT. I said the same exact thing. I’m like you, this wasn’t great because I need to be indexed for these keywords. Here is my list. Now it didn’t accept the CSV file. It was like hey, attach a list of keywords. And it said oh, it didn’t come through, can you attach against? What I did was I just copied and pasted all 85 of those phrases, and then it’s a.

Bradley Sutton:

Andrew’s GPT is very kind. It said thank you for the keyword list. Well, very, um, a lot of good manners here. And it says hey, you know what? I’ve rewritten your listing to maximize keyword indexing while keeping it engaging and con conversion friendly. Blah, blah, blah, blah, blah. And then it spit out a new listing. Now watch this, guys. This is pretty cool.

Bradley Sutton:

The new listing I went ahead and put back into listing builder. And look at this score 534,000. All of and look at this score 534,000. All of a sudden it is now the number one, ranked one, and of the high search volume keywords, the top 10, it had all of them except three. Out of the most relevant keywords, it had like five or four out of 10. And out of the individual 85 keywords, it was now indexed for about 60. So that I mean that one did even better than the Helium 10 listing builder. AI might need to tweak it, but you would expect that because, like I said, Andrew is literally making these a hundred percent tailor-made for what works on Amazon. So all it needed was to get that extra data of what keywords it needs to use that we need to be indexed for. So again, moral of the story here, guys, is 100%, you should be using AI when appropriate for your Amazon business. It can automate a lot of tasks. It can do things that maybe you couldn’t have done, like in graphics and customer communication and research. But there’s a right way to use AI and there’s a wrong way. The right way is relying on AI when the AI has all of the valid information. Right, if I’m telling AI to make an infographic based on this image, I’m literally giving it all the information it needs. Here’s the image, here’s what I want on the infographic. It can handle that.

Bradley Sutton:

But it’s a different story when we’re talking about trying to use it for keyword research. If it doesn’t have the data from Amazon or from Helium 10 about search volume and rankings and stuff, it is not going to be basing the output on accurate information that is highly relevant to Amazon. It’s probably going to be basing off of Google or other data sources, kind of like the old days of before there was a Helium 10 and other tools. You know people are using Google tools to optimize for Amazon, which is not terrible, but it’s not great, okay, because it’s not Amazon. Now, same thing with listings. All right, if you know you can rely on ChatGPT even a custom ChatGPT to make a listing that looks like it’s really optimized for conversion. But at the end of the day, guys, indexing is still about what keywords you have in your listing, and if you’re not inputting anything to ChatGPT, it does not have all of the relevant information that it needs to really write a listing that is going to get you not only conversions, because it’s great copy, but that is going to get you indexed for all the important keywords that can bring you sales. So conclusion again use AI, but make sure you are inputting data from Amazon be it Search Query Performance, be it product opportunity explorer or, of course, obviously the easiest one is Helium 10, and then take that data and use it so that ChatGPT or whatever kind of AI tool that you’re using, it’s going to have all of the information that you need or that it needs in order to give a well-optimized Amazon-centric result.

Bradley Sutton:

All right, so let me know what you guys think in the comments below. How are you using AI for your business? How do you want Helium 10 to maybe invest a little bit more into AI? Let us know. One thing I’m trying to get the team to do is something that can use AI to make virtual bundles, which is a super tedious task to do it manually on Amazon. I’m sure you guys all know about that. That’d be cool if Helium 10 could do something AI related there.

Bradley Sutton:

Maybe it’s time for us to update some of our AIs that are existing. You know that we haven’t updated in the last couple of months. You know AI is constantly changing. You let us know. Let me know in the comments or on my Instagram Serious Sellers podcast or in the Helium 10 members Facebook group. How would you like Helium 10 to integrate AI a little bit more? That could help you with your brands. Let us know and until the next episode, I wish you the best of success, whether using AI or not, whether using Helium 10 or not. Hope you’re crushing it on Amazon.

Enjoy this episode? Be sure to check out our previous episodes for even more content to propel you to Amazon FBA Seller success! And don’t forget to “Like” our Facebook page and subscribe to the podcast on iTunes, Spotify, or wherever you listen to our podcast.

Get snippets from all episodes by following us on Instagram at @SeriousSellersPodcast

Want to absolutely start crushing it on Amazon? Here are few carefully curated resources to get you started:

• Freedom Ticket: Taught by Amazon thought leader Kevin King, get A-Z Amazon strategies and techniques for establishing and solidifying your business.
• Helium 10: 30+ software tools to boost your entire sales pipeline from product research to customer communication and Amazon refund automation. Make running a successful Amazon or Walmart business easier with better data and insights. See what our customers have to say.
• Helium 10 Chrome Extension: Verify your Amazon product idea and validate how lucrative it can be with over a dozen data metrics and profitability estimation.
• SellerTrademarks.com: Trademarks are vital for protecting your Amazon brand from hijackers, and sellertrademarks.com provides a streamlined process for helping you get one.