Episode 59: AI, Discoverability & GEO: Transforming Book Marketing with Shimmr AI

Brooke Dobson is Co-Founder and Chief Revenue Officer of Shimmr AI, a fully automated platform for publishers that creates automated advertising for discovery and sales of books. With expertise in economic forecasting, marketing analytics, and AI technologies, Brooke brings a unique view into AI applications in publishing.

She joined us on the Booksmarts Podcast to discuss how advancements in AI are transforming book marketing, as well as the importance of metadata and Generative Engine Optimization (GEO), and practical strategies publishers can use to boost discoverability, connect with readers, and enhance their marketing through automation.

Learn more about Shimmr, visit their website and be sure to read their recent article—featuring Joshua Tallent: How to Make Your Books AI Discoverable: A Publisher’s Guide to GEO.

Digital Transcript

Joshua Tallent
So this month on the book smarts podcast, I’m excited to talk to Brooke Dobson. Brooke is the co founder and chief revenue officer at shimmer AI, which creates automated advertising for discovery and sales of books. And Brooke and I have known each other for a little while. We’ve been kind of back and forth seeing each other publishing conferences and kind of watching what she’s up to, and very excited to see the work they’re doing at shimmer. So Brooke, thanks for joining me.

Brooke Dobson
Thanks for having me. Joshua, equally excited about the things that you and Firebrand are up to as well. So I’ve really appreciated the kind of collaborations and whatnot, just getting to know you over the last couple of months.

Joshua Tallent
Yeah, it’s been really great. And I, I love the the fact that you guys are really, like, focused in on marketing, focused in on what publishers can do to help their books succeed, because that’s obviously a big thing that we do at Firebrand as well. And there’s, there’s always a need for more, more marketing, better marketing. And with the AI revolution that we’re in the middle of, you know, what do I talk about this way too much on every, almost every podcast that comes up now, but, but it’s important, right? There’s, there’s things that are happening in the industry, and we need to kind of stay abreast of them.

Brooke Dobson
Yeah, that’s, that’s very true. I mean, the way that I see, the kind of moment in time they’re at right now. I’ve spent my entire career in marketing in different forms, you know, whether it be in analytics or in advertising. Now, for the last eight years or so, using AI to help marketing or advertising in various ways, and I’ve seen marketing, and specifically social media advertising, get harder and harder because of what I call kind of the and I call this kind of like the book marketing crunch that we’re in right now. And I think AI actually solves every one of the problems that I have seen getting worse and worse. So those things that I’ve seen are social media fragmentation. So there are so many different places that you can find your audiences now it’s very difficult to know where exactly they’re going to be, and they might be and they might be in multiple places. You might be able to hit them in multiple touch points, which you know is important. So finding them is a challenge. Reader attention and really just consumer attention to ADS is down as well. You might get two seconds from somebody to really catch their attention with an ad because of so much that they’re seeing online. So those two things coupled with each other, and then on top of it, in the book industry in particular, which I’ve spent now the last three years in books, don’t have much profit margin to be able to sustain the type of uncertainty, high waste of traditional marketing and advertising. So it really has to be made much more efficient. And so the way that I see, I see AI being able to impact all of these things. AI in terms of finding your audiences. AI enables you to take your book and basically understand all of the psychological and structural depth of the book, to identify all the all of the people and all of the things that might be that might be of interest to people in that book, and then use that to target audiences. So you can start much broader than before, and use AI to really test and learn what works within that much broader but within the confines of your actual book details. So I think that really helps with the problem of where to find people. AI can also test and learn on lots of different channels and figure out, figure out quickly which ones to turn off and which ones to keep on. So start on Pinterest, LinkedIn and YouTube, and figure out where your audiences are and turn off very quickly the ones that aren’t working, you know. So that kind of cuts down on the waste and the in the in the that high kind of human effort that’s involved. And then it also allows you to advertise emotively, because you can then take that the contents of your book that AI has understood, and allow it to build out creatives for your marketing efforts and specifically for your advertising that are going to catch people within that two second span that we have online. And it’s a much better way of catching them. There’s this concept called system one and system two, if you’re familiar with it, system one being that kind of subconscious you know the the gut feeling that you have. And if you can catch people in that subconscious one or two seconds, you’re much more likely to get them to purchase something. So I think it’s all of those things that AI can now help us to solve with better tools and the way that it can experiment and learn much faster.

Joshua Tallent
Yeah, and that ideal customer profile idea, right? The idea of finding who are the right people for your book, there’s a lot of problems with that because, you know, some some publishers focus way too much on what does the editor think? Or. What does the author even think and less on really? What’s happening like real data, who is actually reading those kinds of books? We’ve seen this before, where a publisher came to us to talk about kind of, you know, figuring out issues with their metadata, and they actually found out that they were most of the people buying their books were not their ideal customer profile. They had built out this whole profile of who was, who was their target market, and it just wasn’t right, and that’s the kind of thing you see. So being able to use AI tools to help you do that work to figure that out, that’s really important, right? We’re talking about taking advantage of opportunities and using tools in the right way to help them benefit the overall visibility and sales of your books.

Brooke Dobson
Yeah, we actually call that. We dubbed a term called reverse discoverability. Actually, Porter Anderson used this term first when he started looking at what my business shimmer is doing. But what reverse discoverability means is I is do an in depth analysis of all of the book contents, and then use that to create hundreds of interest groups, hundreds of keywords that just identify anyone who could possibly be interested in the book, and don’t even worry about demographics or the character traits that you were focused on in the past. Allow the book to find all the potential readers and then learn quickly what to turn off. You know, that’s what AI can do really well. And then it can tell you, then you learn on the back end, who are the demographic profiles of the people purchasing my book, what are their other characteristics, rather than you having to teach it that on the upfront and perhaps miss a lot of people.

Joshua Tallent
Yeah, and that’s obviously if you, if you’ve ever done Facebook advertising or anything else, it’s, I don’t know who to target like. I’m guessing it’s, it’s just a guess. But having having that data is very, very important. So if you think about other areas where AI is having an impact, practicality is a big thing for me. I don’t like telling publishers, oh, maybe try this thing over here, or be, you know, being too vague and kind of but being practical. What are the things that publishers can do? So thinking about that from a publishing perspective, what are some areas where book marketing is really being impacted by AI and especially, what kind of tools are you seeing the publishers are using? Obviously, there’s some experimentation that always is going to happen. It’s not just about. It’s not just about I know exactly what I’m going to do, which tool I’m going to use for this, or for that, or whatever. Sometimes you have to experiment. What are the things you’re seeing publishers take advantage of the tools and the opportunities, the practical advice you might give for someone who’s just trying to figure this stuff out? Yeah.

Brooke Dobson
First of all, there, there are places throughout the marketing workflow that you can use AI. And so it really comes down to kind of what is the easiest or lowest hanging fruit areas for you as a publishing business to get started into and then only keeping going with those things that are actually making an impact on your business’s bottom line. So also that kind of keeps you focused on, like, I like to say, figure out, what are your businesses priorities when it comes to marketing? Use AI there first, because your pilots are probably, even if they’re successful in that they accomplish whatever metrics they were supposed to accomplish on that one thing that you were trying to do within your marketing, if they’re not actually solving a business priority, you’re actually very unlikely to continue that on through your business and scale. So in that, in that sense, it’s almost like a wasted effort the areas that and there’s also tons of tools that you can try within each one. So just to go through kind of a list of the areas that I have seen AI being used the most in marketing. You have research so, I mean, that’s kind of an obvious one, but there are are you can use AI in such an easier way now to research topics. You know, anything that you’re thinking about with your marketing, kind of get ideas for your marketing efforts. Of course, what you’re very familiar with is metadata and SEO. So now there’s also, I mean, I say metadata and SEO. There’s also Geo, which we can maybe come on to later, because that’s a whole nother new topic. I don’t want to get confused right now. But keyword enrichment, you know, picking categories, picking your keywords, you know, all of those things are, are much easier done with AI, and it actually enables you to enhance your metadata. I remember you talking Joshua to me at one point, because you’re much more of an expert in metadata than I am mentioning that it’s not only important that you keep your metadata up to date, but it’s equally as important that you do it at the right time. And so if you have a tool in place that can tell you the right time to update your metadata, then AI can help you do that in a much easier way, so you’re actually not missing the wave that you might be able to ride for that book. So I think that’s pretty important. And then asset production, of course, I think this is probably the place where most publishers have gotten started, is like with their content, you know, so creating mess. Aging, creating your A plus content, creating your blurbs, all of those things you can do with AI, even if it’s not your finished content, that you use your AI output for just giving you different ideas for iterations of it, and then putting your own human touch on top of it, which is often gives it that kind of authentic nature, which I think is pretty important these days, given the fact that so much content is now being created with AI. It’s it’s more obvious now when something as a human layer on top of it, when it’s not fully created by AI. So I think that that could be kind of a competitive advantage that you actually give to your AI content when a lot of other people are just using AI now to create a lot of those things, and then testing and optimization, that’s probably the final one I’ll point out. So before, when you’re creating ads for a book, you might be able to manage creating one or two creatives for your book, and you don’t know whether those are going to going to actually appeal to just one of your audiences or all of your audiences. They might just fall flat in general, and then you waste a lot of money trying to figure out how to make them work, when they might just not be a good fit, or you might not be targeting the right audience. So now, being able to test and iterate quickly and learn quickly from the AI what’s actually working means that you you waste a lot less.

Joshua Tallent
Yeah, those are great. I think the metadata side of things is, obviously, I mean, obviously, I’m a huge metadata nerd. That’s my thing, but it’s, but it’s also one of the most important areas for a publisher, because there’s, there’s so much effort that needs to go into it, but also so little effort that unfortunately does go into it sometimes. And you’ve got a long list of titles, and if you’ve got a backlist, how do you go back and clean that up? And how do you know when the right time is, and all these other things. So let’s, let’s talk about a little bit about Geo, because I do think important topic. We, last month, just released a blog post together on your website that will link in the in the show notes, talking about Geo. But let’s, let’s kind of give a little bit of a top down perspective. What is geo and how is it different from SEO?

Brooke Dobson
So in the most simplistic way to describe this, probably everyone listening to this podcast knows what SEO is, or search engine optimization. That’s basically helping your the people who are searching on Google Search to be able to find your book. You know when they’re searching for for a book, bit of your genre or of your category, but geo is generative engine optimization. So this is about optimizing your metadata so that when people are now more and more searching in chat, GPT, Gemini, you know all of the AI assistance, which they are using a lot more frequently. Now it’s helping, it’s helping those engines to find your book. And unfortunately, just as everything goes, you know, Geo, what it looks for is is different than what SEO looks for, whereas SEO is focused on, you know, keywords and things like that, Geo, the way that people talk to a chat, GPT to a Gemini is very conversational. They’ll give it more information, because generative engines can can take a lot of different things that you want out of a book and find that book without it, without the exact words being found in a certain article. You know, so making your G making your keywords and your metadata more robust now, so that people can find so that the generative engines can find it with that more you know conversational type text is what will help you get found. That’s what geo generally is.

Joshua Tallent
Yeah, it’s more natural language, right? It’s, it’s exactly. It’s looking at data in a different way than a search engine does, because the search engines typically are just indexing content. They’re going to find whatever’s out there. They’re going to follow a bunch of links. They’re going to go to different pages and try to see, oh, how important is this site compared to another? And there’s still some element of that with geo but, but really, it’s, it’s digging deeper into sentiment. It’s digging into context. It’s trying to understand the content as much as, you know, computer can understand, but actually understand it like really think about the the content that it’s seeing. So and this also, you know, comes into into play when you’re talking about things like keywords. You know, we put a lot of effort into Onyx metadata, putting data into your Onyx feed, sending it out into retail stores, into online bookstores, and whatever else. And the thing is, the data that you send in an onyx file is mostly being hidden behind the scenes. It’s hidden in a database somewhere that’s not necessarily going to be accessible to a geo engine, right? A generative engine is not going to see the keywords because they’re not on the actual HTML page. They’re not in the visible content. This is especially true for Amazon, right? Because Amazon uses keywords and their engine, you know, their search engine, and how they how they surface content in the search but they don’t make those keywords available on the actual HTML page, so it never gets seen by chat, GPT or Gmail. Or whoever else. So when you’re looking at expanding your visibility for your books in a geo environment, what do you think are some of the most important things that a publisher can do? What can they do to make sure their book is not lost in the shuffle?

Brooke Dobson
So there are, there are kind of some steps that you can take, and the article that you’re going to post will be helpful, because that does go through the step by step that that your team and our team kind of put together. But first you’re going to want to start with metadata precision. So as you mentioned, like plain language fields, you want to make sure you have the genre, the setting, the themes, the audience, kind of the more robust information that the geo engines are going to need. You want to write AI friendly descriptions. So this means that your descriptions are now not just kind of, you know, jargony or trying to grab attention, but actually describe the plot, the tone the character hooks, you know, you want to avoid, kind of that vague hype that is not going to be found in the generative engines you want to add supporting thematic content. So blog think about like blog, blog posts, your sub stacks, you know, guides, interviews, anything that you’re doing basically gives AI more and more context in order to be able to find your book. So the more you give it. Because again, think about how the AI engines actually go out there, and they find they pull lots of different pieces together, you know. So the more content that you have out there that helps link to the things that people are searching for, the more likely your book is to come up and be found. And it is really important. I saw a stat the other day that said something like the traffic that’s going to publishers websites could decrease up to 50% if you don’t actually focus on enhancing or putting in place a geo strategy. So the implications of the implications of doing this now and kind of getting ahead of everybody else, because this isn’t something that that all the publishers and authors are jumping right into, so kind of being on the front line now can get you in a competitive advantage, but not doing it can can give you a serious competitive disadvantage based on some of these stats.

Joshua Tallent
Yeah, and I think one of the biggest things is to remember that more content equals more visibility. For a search engine, a generative engine is going to look for every single bit of data it can come up with. It’s not just looking at little things like keywords. It’s looking at, do I have additional content, and this is where your website is so much more powerful as a publisher than, say, Amazon will be, or Barnes and Noble or bookshop, or whoever, because you have so much more control over your website. You can put in character bios, you can put in descriptions of other content in the book. You put in extra tables. You can put in other content, PDF documents and whatever else. Ai engines can see all of that, and they can use all of that. And the more it’s linked to the book itself, the more it connects to what that book or what those books are about. And that’s that’s really important. It’s one of those things I’ve been hammering on for a long time, is more content is better, because more content gives more opportunity for more visibility, and now we’re seeing that come to fruition in a way we didn’t even expect 10 years ago. So a book excerpt in your data, a book excerpt on your website as plain text, as HTML text, not just hidden in a PDF. All the PDFs are fine, but you want to you want to enhance that data. You want to build on what you already have and take the opportunity to kind of back into that. Do it? Do it further back in your process. Talk to your editorial team and engage with that. Talk to the author and engage with that. Look at some of the stuff that’s happening from independent publishers, independent authors who were kind of really quick and quick and nimble on their face, working in this stuff, you know, there’s a real benefit to a publisher and saying, How can we be more nimble? How can we adjust more quickly? How can we add more content? And what do we do to expand on the visibility of our content by putting more out there?

Brooke Dobson
Absolutely, you know, and you don’t have to start with the whole catalog at once. It’s actually better not to, you know, it’s better to kind of test and see how the things that you’re doing are going, you know, choose a small number of titles and update it and then compare. Because then you can, you can test it easily yourself. You can start going into the chat gpts, you know, once you think you’ve optimized your geo strategy for your some of your titles, you can go in and just search, as you would a normal consumer, you know, in go in Gemini, go in every single one of the chat, you know, the AI assistants, and try to test it and find your books and see if it’s working, you know, and then being able to do like an AB test, like that. So then do the same type of searching for the books that you haven’t enhanced the geo content. So then you can see, does it seem to be working? And I can tell you, actually, Joshua, we are, this isn’t about specifically books, but our marketing director at shimmer James, has spent the last month or so enhancing, or actually, really just putting into place shimmers geo strategy, so that when people are searching for, you know, advertising for books, or automated advertising, or AI for you know. Books, then they find shimmer. And the last, in the last two weeks, I’ve had three or four demos that have come inbound to us that have said that they look I always ask, you know, how did you come across shimmer? And I’ve never once heard them say a Google search, but three or four of them have in the last two weeks, said, I just went into chat GPT, you know, that’s what they’re doing these days, you know? So I’m like, okay, the geo strategy is working. We just put it into place.

Joshua Tallent
That’s great. And I’d like to actually kind of dig into a little bit of what shimmer is up to. So can you explain what shimmer does and how you guys are helping publishers with this part of their of their marketing strategy?

Brooke Dobson
Sure, sure. So you mentioned, like, the importance of the backlist earlier. We feel the same. We we are. Our founding team came from a background of traditional marketing and advertising. Then we did a previous AI business in the space, space of brand management together about eight years ago. So we’ve been working in technology and using AI to solve problems in marketing, long before chat GPT or generative AI was a thing, long before any of us were even talking about AI. And when we decided to do when generative AI came onto the scenes in November of 2022 I think it was, it gave us an opportunity to do much more with AI than we were able to do prior to the generative AI. And so we decided we wanted to do autonomous advertising. You know, it kind of brought all of our skills together, and it was now something that was able to be done. You could build basically an AI agent that does all of your advertising for you by just having access to the product itself. And we launched into books. And we only work in books, because what we felt was there are so there’s such an opportunity here, and that 95% or so of books go unpromoted and unconnected with their audiences for the majority of their lifetime. And that has a lot to do with some of the problems I mentioned earlier when I talked about the book marketing crunch. You know the fact that book profit margins can’t sustain wasted marketing and advertising for a long period of time, and so much of it is wasted, so it just doesn’t give you the ability to do it for all of your books. You’re kind of chasing the best new book, you know, when you’re a publisher. So we launched into book publishing, and what we do has really gained a lot of traction quickly. Essentially, we only require the electronic file of the book. So because we know, you know, if you’re going to be able to liberate your entire backlist, there’s no one in the publisher or in their marketing team that is dedicated to their backlist fully so there’s not a lot of people there to do anything. So you needed a fully autonomous product in order to be able to truly serve the backlist and the front list. So essentially, we need the electronic file of the book. It goes through three machines. Our strategists will analyze the book, build out what we call book DNA. Book DNA is the in depth, psychological and structural understanding of the book that then allows the generator, which is the second engine, to generate the ads that are that have emotional, you know, qualities to them that really pull people in very quickly. And it also identifies the target audiences to match to the book DNA. It then passes that information to our third engine, which is our Deployer. The deployer sends the ads out to the people online who we want to show those ads to. So through meta, think about all of the social media channels, and then the whole entire system is self learning and self optimizing. So we then use the AI to do that AB experimentation that I was mentioning, you know. So test lots of different target audiences, keywords and interest groups to target those audiences. Creative executions for the ads, different online placements, basically home in on the ideal advertising recipe for the book. And then say, media spend, put go here, yeah, turn off everything that’s not working, you know, so that you can really reduce that uncertainty. And we’ve seen the campaigns generating really great metrics. We do, like, six to 10 times better click through rates than the typical like marketing teams for books, and it has a lot to do with it has a lot to do with the fact that we’re showing books. We’re kind of bringing the narrative out of the book, you know, rather than just showing the cover of the book alone, and we do that both in image and also in text. You know, you can pull out kind of the hook of that book and and also target the right audiences. So it generates really strong metrics and just allows people, marketing teams to advertise a lot more books connect them to their readers that they weren’t able to do before.

Joshua Tallent
Yeah, that’s great. And that’s that’s pretty, pretty close in some ways, to this kind of stuff we’re doing on the metadata side with our flywheel tool, because flywheel is able to figure out what is the best time to put out product data updates to a book into the supply chain. We’re not doing the advertising side. You guys are doing that and doing doing the kind of front facing advertising, but you have to update your metadata too. You have to make sure that the product data that’s out there is being kept in sync with what with all the changes, there’s so much cool stuff you can do when you put a little bit of machine learning behind something, you start to see, let the system think, let the system figure. Are out for you and don’t, don’t rely on your own brain to necessarily know all the you know, be able to process all the different data points. There’s a lot of cool stuff you can do when you have those, those tools in place. So thank you very much.

Brooke Dobson
Yeah, yeah, yeah. And actually, Joshua, I’m still very excited. I know I’ve said this to you, but I still believe that there is a great collaboration between these two products. Because you know, the moment that you enhance that, you know you need to enhance the metadata, and you do so, and then you then get the books on to advertising that already are hot. You know, it’s like just getting the hot potatoes on there. I feel that there’s something great that’s going to happen between those two, between our two companies at some point here.

Joshua Tallent
We should talk more about that. Yeah, yeah. Well, thanks, Brooke. I really appreciate you joining us today. I really appreciate all the work that you’re doing. And you know, this is stuff that I mean, we’re all struggling with it. We’re trying to figure it out. Everybody feels kind of overwhelmed, I think sometimes by the number of AI tools and all the different which bot do I use and how do I use it? And don’t get overwhelmed, just start using right just take advantage of it. Do whatever you can. A little bit is better than nothing, and at least a little bit, you start to learn as you start to use it. So appreciate your insights and the work that you’re doing at shimmer, where can people find more about what you’re doing online?

Brooke Dobson
So you can go to our website, which I’m sure Joshua, you’ll put somewhere, but it’s shimmer.ai. So everything about our company is there, and I also you can find me on LinkedIn as well.

Joshua Tallent
Well, that’s it for this episode of the book smarts podcast. If you like what you heard, we really appreciate it when you take some time and share it with your friends and colleagues. If you have any questions or suggestions or feedback about the show. You can reach me at fire Joshua@Firebrandtech.com and we also appreciate any reviews and ratings on all the podcast systems out there. Thanks for joining us today and getting smarter about your books.