How to Send Emails that Stand Out with R.J. Talyor

Marketing is starting to look the same everywhere, and AI is accelerating the slide toward safe, average choices. But that's not what good marketing is. In this episode, we dig into what the data says about email performance and what we can change right now to make our messages connect on a human level. 

I’m joined by RJ Talyor, Founder and CEO of Backstroke, who’s spent decades in digital marketing and email, and built AI tools that predict what people click. He shares insight on what's really working in email marketing. 

Download the FREE "What Do You Do?" Formula Guide here.

Transcript

Copy Transcript Copied

Janice Hostager: [00:00:00] Have you noticed how much marketing has started to feel all the same? The same graphics, the same Canva templates, the same AI-written headlines, the same colors, layouts, calls to action. And the word "actually" is a red flag that screams AI to me. I don't know. 

Janice Hostager: Maybe it's just me. Anyway, if you scroll through your inbox, after a while, it all s-starts to blur together. But here's the funny part. Most of us aren't really trying to copy each other. We're just kind of following the same best practices. We read the same blogs, watch the tutorials, use the templates, and then we let AI help us write maybe a little faster. 

Janice Hostager: Before we know it, our marketing looks a whole lot like everybody else's, and it all becomes plain vanilla. So how do you stand out without throwing away everything that works? That's exactly what we're talking about [00:01:00] today. My guest is RJ Talyor, CEO of Backstroke. His company has analyzed more than 20,000 retail marketing emails to understand what actually grabs attention and what causes brands to disappear into the sea of sameness. 

Janice Hostager: Today, we're digging in why algorithms often reward conformity and what small business owners can do to build marketing that's both effective and unmistakably their own. If you've ever wondered how to make your emails or your brand or really any part of your marketing more memorable, you're gonna love our conversation. 

Janice Hostager: Here's my talk with RJ. 

Janice Hostager: welcome to My Weekly Marketing 

R.J. Talyor: Hey, thanks for having me. Appreciate it 

Janice Hostager: So before I dive in, usually I like to jump into questions right away, but I really think that your background is super relevant to the discussion that we're having today. So tell us a little bit about, how you got into this topic because I think it sets up our conversation nicely. 

R.J. Talyor: Yeah, I mean, I've been [00:02:00] in digital marketing in, in the software world for twenty-three years now. I got my start at ExactTarget. I led the email and mobile teams at Salesforce as a part of the Salesforce Marketing Cloud, and then I really got into the creative AI space in 2016. I started a business that was called Pattern Eighty-Nine, and it would predict what you would click on, on Instagram or Facebook. 

R.J. Talyor: And so it would predict, that women who live in Austin who wear black shirts like, you know, blue couches, and that men who have brown shirts who live in Indianapolis like, certain other dimensions. And, it was early days of AI, and it helped us to use computer vision, and tagging to extract features from large datasets to then predict what we would actually click on. 

R.J. Talyor: I sold that business to Shutterstock in 2021 and then, started tooling around with some of these generative capabilities to start Backstroke. And Backstroke's my company now that predicts what you'll click on, based on personalization within email campaigns. [00:03:00] And, so that's kind of the quick professional background. 

R.J. Talyor: In addition, I have a, an English degree, and I have a master's in creative writing. So I come at all of these things from kind of two angles. One is from the technology capability side, and with an MFA, I come at it from the creative side, and it's kind of this weird intersection that we all are at from, like, a human perspective and it's kind of a fun place to be, I think, to experiment and learn. 

Janice Hostager: It really is. I mean, that's kind of the same background that I have too. I have a degree in art and also English, and then I also have, a background in running an agency, creative agency, and then now I'm in marketing strategy. And of course, I think I've really tried to get in there and really embrace AI, but I'm starting to see the issues with all of this, and I'm starting to pick up on a lot of it, when I see it in other places. So I think that's what makes this, such an important conversation to have. Now you have a tool that you evaluated like 20,000 [00:04:00] plus brand emails, right? 

R.J. Talyor: Yeah. Yep. Yep. 

Janice Hostager: right? So, 

R.J. Talyor: Yeah 

Janice Hostager: when you opened the data for the first time, what was the thing that made you say, "Oh gosh, I see an issue here"? 

R.J. Talyor: Yeah. Well, yeah, so, back five years ago, I started signing up for every email marketing program I could find to build the data set that we could then analyze these tools. And so we have millions and millions of emails now across all these different, retail and e-com brands. And the thing that's shocking is the sameness of it all. 

R.J. Talyor: And I, like I-- it, it's, it's so, as humans we're, normalizing towards best practice, and there's a lot of mimicry out there. We look at the leader, and then we try to make our content look like the leader because we want it to feel more aesthetic or more, luxe or more expensive or more sporty or raw or whatever it is. 

R.J. Talyor: And the true, like, human voices get kinda squished out because as humans, we are attracted [00:05:00] towards, looking the same or looking on trend, right? Like, we wanna sound trendy or we wanna sound like the leader, or we wanna sound like each other. And so the sameness was actually really what was wild. And, you know, everything, like in the email data set from like, the average number s- of words in a subject line is six, six words. 

R.J. Talyor: There are lots of common words that are always capitalized, or there are seven most common emojis that are being used. And it's, you know, the u- the lack of uniqueness in the email. If you think about a structure of an email, it typically is the first four hundred pixels has, a face or a body, and then that's the hero with a Shop Now button followed by a four grid or a six grid of content. 

R.J. Talyor: And so like these templates and subject lines and colors, everything almost like looks so similar within a single email. And then if you look at the emails next to each other, like, the pattern is very similar as well, and the, the machines pick up on it, and they identify that, like, [00:06:00] when you introduce the, summer line, it's an introductory email, then it's feature one, feature two, colorway, then it's feature two, feature three, specific down, like zooming into the fabric. 

R.J. Talyor: And then we move into a sale frequency where introduction of the sale, the sale got discounted to ten percent, twenty-five percent, then it's clearance and clearance on clearance. So it's like a set of around fourteen to sixteen emails that is within one of these collections that's then repeated throughout the year. 

R.J. Talyor: So sameness was what the shocking thing was, 

Janice Hostager: Yeah. 

R.J. Talyor: the data. 

Janice Hostager: Yeah, that's really interesting. And I notice this a lot, especially with newer businesses because especially the people like my listeners, I mean, we're starting out. We're, a little unsure of what everything should look like, so we immediately, grab onto somebody who's doing it really well, right? 

R.J. Talyor: Uh-huh. Yeah 

Janice Hostager: and so then we think, "Okay, if I could just do it like that, then I'll, be more successful in what I do." And I noticed this too working in advertising years ago and [00:07:00] marketing, it really is the part that people just look right past. 

R.J. Talyor: Yeah 

Janice Hostager: It doesn't stand out to them because they're so used to it. And we really need something that's gonna stand out and look different, which, is obviously, what works. Sometimes it works. I won't say always. But 

R.J. Talyor: Yeah 

Janice Hostager: that's, that, that gets the attention, and that's... 

Janice Hostager: The attention is that thing that we're real all trying to get. 

R.J. Talyor: Yeah, yeah 

Janice Hostager: But where is that sameness coming from? Is it that we kind of all look to the leaders in our industry and people that we think are doing well and try and copy that? Is it from the AI models themselves, producing similar outputs? 

Janice Hostager: Is it, maybe the prompts that everybody is asking from AI? 

R.J. Talyor: Well, y- y-- there's kinda two answers to the question. One is like, I mean, I'm not an anthropologist at all. Like, I, but, like, you know, as humans, we're pack animals, and we do things to fit in, right? We wanna look in a certain way or act in a certain way so we're not rejected. And, like, at the primal level, you don't get cast out and starve and die or whatever. 

R.J. Talyor: And we wanna l-- you know, [00:08:00] we wanna look and feel the same. And so-- Or feel s- you know, similar enough that we're, expected and, familiar. So there's that. You know, wild ideas are rejected often. You know, like in the ad agency world, everybody says, "I want something unique," and then you show them something unique, and they're like, "No, I want it to look like Nike." 

R.J. Talyor: Or, you know, and so... 

Janice Hostager: Yeah 

R.J. Talyor: or whatever the case is. So, I really admire trailblazers and unique people and weird things and people in-- for, for that reason. Like, I love-- I could tell you all the weird things I like. You know, like, but I love that. And, but it's dangerous. It's risky. And so oftentimes we, you know, we, we kind of, normalize, right? 

R.J. Talyor: We go to the norm, and we wanna look in a certain way, act in a certain way. And, you know, the AI models are predictive models. These LLMs, they are predictive, and, the predictive is looking for what are you gonna click on, what are you, what's most likely, what are the patterns? And it's trained on these large datasets, that have a sameness to them. 

R.J. Talyor: And so it's looking to h- [00:09:00] again, normalize, the data and say, "Hey, here's what a, quote, best practice is." Like, I'm so guilty, maybe you are too, of writing lots and lots of blogs from, like, ten years ago. Here are the ten best practices for retailers. Here are the ten best practices for subject lines, ten be- Like, and what we're doing is we're saying, "This is the average." 

R.J. Talyor: Best practices are average. They're the things that are, like, what are expected. And so that's effectively what all these LLMs are doing, is looking for the best practice and then recommending that. And so between the combination of sort of our innate need to fit in, in combination with the LLMs just reinforcing it, we end up with a lot of sameness, a lot of khakis, a lot of, you know, just a lot of, things that look the same, whether it's, you know, like, the all white kitchen, the Lululemon, leggings look, the, you know, clothes that I'm wearing or, you know, the emails and marketing campaigns that we're sending. 

R.J. Talyor: It's, [00:10:00] it's, kind of a double-headed monster there. 

Janice Hostager: Yeah. And you're right, we all fall into that to some 

R.J. Talyor: Yeah. 

Janice Hostager: in some part of our life for sure. 

R.J. Talyor: Yeah. 

Janice Hostager: the weirdo, right? 

R.J. Talyor: Right. 

Janice Hostager: wants to be a weirdo. 

R.J. Talyor: Yeah, totally. Yeah 

Janice Hostager: so the people that are listening to this, they're using ChatGPT or Claude or to help them write emails. When an LM looks at the data, do they at what's successful or do they look at the averages? 

R.J. Talyor: Well, they're looking at what's common. So the LLMs don't have the performance data in them. They're looking at what is most commonly used. So I mean, that's the whole pitch for my company Backstroke, which is we can predict what's gonna perform, looking at your performance against a larger dataset of thousands of, other retailers and millions of, email campaigns so we can understand what, is ver- what is on brand, as well as differentiated, but then also performant. 

R.J. Talyor: Because, you know, as a marketer, you're gonna get fired [00:11:00] if everything's like weird but not performant. 

Janice Hostager: Right., Yeah 

R.J. Talyor: if it's, different and unique and on brand, and it performs, then of course you've got a, a really nice winner there. So, yeah. So to answer your question, the LLMs just look for what's most common. 

Janice Hostager: Okay, interesting. 'Cause I tend to trust them thinking like there's some all-knowing \ entity out there, like 

R.J. Talyor: Yeah. Yeah 

Janice Hostager: really well, so I should trust them even though my gut's telling me that's not the right approach. 

R.J. Talyor: Yeah 

Janice Hostager: 'cause I kinda go back and forth with them, for like more bigger decisions and less like the writing and that sort of thing. 

Janice Hostager: But, that's really interesting. So they're, designed to just move toward the average 

R.J. Talyor: Yeah. Yeah. And then like, and I, I wanna urge like that's okay for, for some things. Like that's okay. You can't like make everything a masterpiece, you know? Like you'll go out of business trying to make everything a masterpiece. So, you know, what I talk to our clients about is like, all right, what are like the 70, 80% of the emails or the campaigns that you can allow the machine to generate for you? 

R.J. Talyor: [00:12:00] Do that, and then use your human time to actually create something magnificent, something unique, something that stands out. And like that's the whole point of a tool that can automate or, you know, perform a task for you. Where I think people make mistakes is they say, "Oh, now I'm gonna offload this whole thing, my entire marketing strategy to the LLM so that I can go do something more important." 

R.J. Talyor: And if you're a marketer, I'm not sure what's more important than literally 

Janice Hostager: Getting 

R.J. Talyor: your job. 

Janice Hostager: Yeah. 

R.J. Talyor: you end up with averages. And I think that we're just at the start of this, because the problem is that the data that's been collected and normalized by these LLMs is from a vast cam-- like set of data. But now we're actually seeing AI-generated content training the machine, you know? 

R.J. Talyor: So it's like reinforcement learning is coming from itself, and so it's just becoming more and more and more and more and more and more normalized. So it's, 

Janice Hostager: Yeah 

R.J. Talyor: the opportunity is for humans, of course, to then create something net new, right? Bringing in, bringing in some dissonance [00:13:00] or uniqueness, which I, which I think is cool. 

R.J. Talyor: So 

Janice Hostager: Yeah, I actually listened to a webinar yesterday by copywriter Laura Belgrade, and of course, she's very creative and does amazing things with her writing. But she was, comparing just a sentence written by AI, which was a decent sentence. It was grammatically correct. It seemed good 

R.J. Talyor: Yeah 

Janice Hostager: that she wrote as a human. 

R.J. Talyor: Yeah 

Janice Hostager: side by side like that, it was really amazing to me because that human element, it c- I think AI can come close in some respects, 

R.J. Talyor: Yeah. 

Janice Hostager: just, it's just not, it's not there yet. And, 

R.J. Talyor: Yeah 

Janice Hostager: you said, I'm afraid it's gonna get worse, but... 

R.J. Talyor: Yeah, me too. 

Janice Hostager: S- 

R.J. Talyor: Yeah 

Janice Hostager: show us the contrast. 

Janice Hostager: What does, like a top-performing email look like now versus a middle-of-the-pack one i- in your data? What is there one big thing or is it a bunch of small things stacked up? 

R.J. Talyor: Well, so for the, the biggest thing that's a difference is the, the hero imagery. So, it would depend on the [00:14:00] individual brand and your subscribers, but 46% of your clicks come from that hero image. So if you think of the first 406-- 400 to 600 pixels, sometimes you'll think, oh, it's above the fold, but it's actually the 406-- four to 600 pixels. 

R.J. Talyor: And within that, it's, , a simple direct m-message with a simple button. There has to be a button, not like a specific call to action. If for some reason, like if there's no button, it doesn't work as well. I mean, but,, the-- that's, that's where the winning comes from. Um, and you could then the rest of the... 

R.J. Talyor: Because that's where the bulk of the clicks come from, right? So if you think about how can I make this part of my email overperform, then that's ultimately the winner. Now there's all sorts of different variations of that, whether or not you include like a, a header bar or not that matches your website, for example. 

R.J. Talyor: Or, we see, for example, like, like a lot of women, tend to scroll down all the way to the bottom of the email on their mobile [00:15:00] device and then click on social icons. We see that there's actually a, a lot of young people who will, click in the first 400 pixels of an email, specifically on mobile devices, but then will actually convert on a desktop. 

R.J. Talyor: So it, it, it's all the classic it depends, but if you think about the first 400 to 600 pixels as the place to really drive the performance, that's where you're gonna get that's where we see the best campaigns winning. 

Janice Hostager: Okay. Okay. So it's more that put your call to action toward the top is, 

R.J. Talyor: Yep. Oh, yeah 

Janice Hostager: of the lessons for sure. 

R.J. Talyor: Yeah 

Janice Hostager: uh, because I was actually just changing emails this morning where I put... I moved a CTA up. 

R.J. Talyor: Yeah 

Janice Hostager: now your study is on retailers, right? 

R.J. Talyor: Correct. 

Janice Hostager: does that, any of that translate, do you think, to service-based businesses, the same kind of thing? 

Janice Hostager: Or are the retail buyers so different from somebody that's gonna buy a service? 

R.J. Talyor: Well, I think that people who are buying p- like products are, are in a sh- like a different shopping mode. It's almost like they're just [00:16:00] looking at a circular in a way. You know, some brands have done a better job of creating a relationship where they're telling stories about their clothing or people who are using their clothing out in the world, or the fabric or the, you know, the, the kind of the legacy behind it. 

R.J. Talyor: On the service-based business stuff, I think there's a lot more opportunity for communicating, , information that, is valuable information that establishes a service organization as expert. So I think that there is more content that they can consume. I don't have the data, but, it makes the, like how do we communicate the hook in the first 400 to 600 pixels and then encourage them to, scroll down and even stay within the email to, learn something or learn about the expertise of the service provider before jumping them off to do whatever's next. 

R.J. Talyor: The call to action is gonna vary. Within a retail environment the, you know, the call to action is buy now. You know, versus maybe a service-based organization, it might be to buy now, but it's probably more like a, a set of [00:17:00] emails that is gonna communicate that message over time. 

Janice Hostager: Mm-hmm. Does that translate to like informational or like I send out a weekly email just with information. There's no CTA on that one typically. same kind of thing or you probably don't have data on that either 

R.J. Talyor: I, I, I don't have, I, I don't have data, but it's inter- you know, I've, I've literally been in email since 2004, so it's, it is the, the case of like building up that expertise, making sure that you're providing something valuable, versus like just driving the click in the retail world. 

Janice Hostager: Gotcha. 

R.J. Talyor: Yeah. 

Janice Hostager: Give us a best practice. So you talked about best practices not necessarily being on track. What's something that we've heard over and over again that your data set is actually hurting performance now? 

R.J. Talyor: Ooh. Let me think about a really good one there 

R.J. Talyor: I would say frequency. We've heard for the longest time that you need to be in your, your subscribers inboxes, three to five times a week, something like that, otherwise they forget about you., About [00:18:00] seventy-seven percent of purchases come from the first week, or conversions come from the first week, and then that, that tail gets really, really narrow. 

R.J. Talyor: So the-- so, it's always been that email's a cheap way to just ping people over and over and over again., You actually end up destroying your deliverability, destroying your overall conversion rate, and then just tiring out the customer if you're communicating to them that many times. It used to be, "Hey, I just wanna be in the inbox. 

R.J. Talyor: I wanna be in the inbox. I wanna whatever." Now you need to really be strategic about who and when, but that frequency component I think is, overlooked, from a best practice perspective, and the data shows it 

Janice Hostager: That is really interesting because just yesterday, I have a colleague that focuses on email marketing, and he put out a video on LinkedIn yesterday saying that the data shows that There was some big study, and I can't remember who did it, maybe Adobe or maybe not, I'm not sure about that, that said that the best thing to do is to email every day of the week, five days, five times a week. So it was just-- It actually [00:19:00] contradicted 

R.J. Talyor: Mm-hmm. Mm-hmm. 

Janice Hostager: said, so that's really interesting. And I pushed back on that a little bit. I said, " when clients have tried that, it tends to backfire. I get a lot of unsubscribes." So, you 

R.J. Talyor: You... Yeah. Yeah, I mean, I would imagine it's business dependent, but like, 

Janice Hostager: yeah 

R.J. Talyor: I mean, I'm not buying shirts every day of the week or I'm not s- you know, like I, I think that's dangerous. I think that my guess is that that would be a short-term win. But because of the priority inbox as well as deliverability issues, I think you're gonna end up tanking your overall email performance in the mid to long term. 

R.J. Talyor: And then I guess the only other, the only other thing I would say is if you have an unlimited budget to get new subs, then sure, go for it. 

Janice Hostager: Yeah 

R.J. Talyor: and y- you know, seriously, like if you can just keep pumping money into the acquisition machine and getting new subscribers on your list, that's great, but you're gonna just, you're gonna annoy/get your whole inbox, or you're gonna get out of people's inboxes, um, over time with that strategy. 

Janice Hostager: [00:20:00] Burn 'em out. Yeah, absolutely. 

R.J. Talyor: Yeah. 

Janice Hostager: So if I'm a coach or consultant and I'm following all the best practices, so I write short subject lines, use one call to action toward the top, 

R.J. Talyor: Yeah 

Janice Hostager: i'm just trying to understand how you tell the difference between a rule that's actually working for your business and versus one that's just, a best practice that everybody 

R.J. Talyor: Yeah. 

Janice Hostager: do.  

R.J. Talyor: listen, 

R.J. Talyor: yeah, I, I would split test everything because the email service providers have made it really easy to do an A/B test, and if you can get in the discipline of A/B testing every campaign and just making one tweak, right? Like, you know, six-word subject line always works. 

R.J. Talyor: Well, maybe it doesn't for you. I don't know. So if you can just make a literally a small change like that, each one you're gonna learn. I mean, if you do five campaigns a week or three campaigns a week and make every single one of them a slight tweak on the A/B test, you're gonna learn what works best for your brand and over time. 

R.J. Talyor: So A/B testing is the best way to, to validate a best practice for you or to learn. The other [00:21:00] thing that I find is that customers often say things like, "Oh, we A/B tested that and it worked." And then you're like: "Well, when did you A/B test that?" Like, "Well, I A/B tested it in December." Oh, okay. 

R.J. Talyor: Well, in December, email volume spikes around 50 to 100%, depending on the day, and it may have worked in December or may not have worked in December, but now it's, you know, a different month, different quarter. It's actually a whole different environment. So constantly A/B testing and testing your assumptions is the way to, to win. 

R.J. Talyor: It does add a step, right? But it's small, and it's almost like you're kinda lured into A/B testing or like to your own best practices. So it's like, oh, we used to test this. I've also heard customers tell me things like, "Oh, Sarah A/B tested that last year." And I'm like: "Who's Sarah? I haven't heard about Sarah." 

R.J. Talyor: And they're like, "Well, Sarah left the business last year, but she's-- she did the A/B test, presented it to the exec team, and that's why we always do the thing," right? And so, like some legacy B- test but done by somebody else, has been accepted as an [00:22:00] internal best practice. So those are kind of some danger points on A/B testing. 

R.J. Talyor: But if you're just in a constant process of trying something new each week, for every campaign, you'll, like you'll validate or invalidate those to find what's, where some pockets of opportunity. 

Janice Hostager: Right. One caution though, I would say on split testing is to make sure you have a big enough list. 

R.J. Talyor: Yeah 

Janice Hostager: if you're just starting out and you have 500 people on your list, 

R.J. Talyor: Yeah  

Janice Hostager: The 

R.J. Talyor: significance. 

Janice Hostager: gonna be skewed. What's that? 

R.J. Talyor: Yeah. Yeah. I mean, like there's, there's, you know, you, you wanna get to statistical significance on some of these recommendations. Th- you know, and you'll get some directional data, but you're right, if you only have 1,000 people on a list, that might be a little, you know, a few, a few clicks can, sway you one way or the other. 

Janice Hostager: Yeah 

R.J. Talyor: but, yeah. So being aware of, of, small data sample is, is definitely a recommendation 

Janice Hostager: So what's something that every business owner, regardless of their size, could change in their next email to make it feel less generic? Would it be don't use AI, use AI to clean it up, [00:23:00] or, you know, start with AI and then edit it? Or what is the rule that they could use to stand out? 

R.J. Talyor: Well, like next email, tell a story. 

Janice Hostager: Mm-hmm. 

R.J. Talyor: Like, like put someone's face in the email and tell like, like, "Here's RJ. Here's why he started the business." Like, "Here's, here's his story." Like AI is not gonna make up the story about RJ Talier, you know? 

Janice Hostager: Right 

R.J. Talyor: go human, there or tell a story about a customer, one of your customers that's been successful, a case study or, a human interest thing. 

R.J. Talyor: Like that is-- it'll take a little more time 'cause you've gotta interview the person, or you've gotta ask for their permission to show their face. But like that's-- I think that's the... I, I hate to word-- use that word best practice, but like that's the way to go. Tell a story, tell it with emotion. Tell about a memory, a dream, something that makes us really human and connected. 

R.J. Talyor: Like that is differentiated and 

Janice Hostager: 100%. 

R.J. Talyor: stand out. Use their name, you know? Put their name in the subject [00:24:00] line. This is RJ's story or whatever. 

Janice Hostager: . And don't use an AI image. We were just talking before we started recording today about, an ad that came out yesterday 

R.J. Talyor: Yeah. 

Janice Hostager: an error on it, right? 

R.J. Talyor: Yeah, well, it, it's fascinating. I, I follow all this stuff to f- um, you know, the AI-generated content, and the big criticism that marketers have is, you know, AI is slop and everybody likes to- Mm-hmm ... you know, AI is slop. And what's-- you know, if you're familiar with the Turing test, you know, Alan Turing, he was the scientist who cracked the Nazi code, and he's responsible probably for me sitting here, you know, and ending World War II. 

R.J. Talyor: But, the Turing test is what helps us underst- is, is the, the language that's used is like, can you determine whether an AI, whe-whether content was AI-generated or it was human-generated? And all these models have now passed the Turing test. It's hard to discern whether an AI or human has created the content or written the email or, created the image. 

R.J. Talyor: And just yesterday, I saw this article that was blasting REI, [00:25:00] the outdoor retailer, I love REI, 

Janice Hostager: Mm-hmm. 

R.J. Talyor: for an ad that they ran on Facebook that featured an image of a bicycle, like a 10-speed bicycle, that had two sets of handlebars on it. And so it was clearly an AI-generated image. And it's a kind of a weird blame game because REI is saying, "We were automatically enrolled in Meta's AI, like beta test." 

R.J. Talyor: And then, Facebook is saying, "No, REI knew that they were in this." And so there's an interesting finger-pointing, but ultimately, like who is responsible for this? Well, somebody didn't look or approve or QA the created image before it went live. And you know, that, that's, that, that's the risk, of course, right? 

Janice Hostager: Yeah. 

R.J. Talyor: When you have like an outdoor audience who values things like grit and extreme sports and stuff like that to get it wrong with an AI-generated image that has two [00:26:00] sets of handlebars and a weird bike, like, you know, it doesn't look good. But on the flip side, as a vendor, Meta doesn't look great for creating that content. 

R.J. Talyor: So yeah. 

Janice Hostager: Well, on Meta ads now, it will say, "Can we just choose images for you?" Or, "Can we, you know, enhance 

R.J. Talyor: Yeah. 

Janice Hostager: for you?" 

R.J. Talyor: Yeah 

Janice Hostager: I've always been hesitant to say 

R.J. Talyor: Yeah 

Janice Hostager: No, I'm not... no. 

R.J. Talyor: Yeah. 

Janice Hostager: images, 

R.J. Talyor: Yeah 

Janice Hostager: So I wonder if something like that happened, because you wouldn't necessarily even see that before it was released, 

R.J. Talyor: Right 

Janice Hostager: a company like REI who's 

R.J. Talyor: Yeah. Yeah. 

Janice Hostager: really interesting. Hmm. 

R.J. Talyor: Well, you, you-- so here, here's, here's the thing where, like, if you ask anybody, like a human, or like if you ask anybody like, "Hey, I'd like to provide you with hyper-personalized recommendations to you, be-- that, you're gonna love, and we're gonna find the right content," most people will be like, "Yeah, I'd love that." 

R.J. Talyor: Right? Like, I'd love for REI or another brand to say like, "Hey, [00:27:00] RJ, you-- we know you'd love this shirt in this size at this price point, that has these features in it that fits really well," 'cause I don't really wanna shop. I don't really like shopping, you know? And I'd be like, "Yeah, that'd be great," like if it actually worked. 

R.J. Talyor: And, you know, that's what we've always talked about as marketers is like hyper-personalization. But to execute hyper-personalization, you have to let the AI do its thing, right? And so create a bunch of ads, like a bunch of content that has all the different variations of the shirt on, all the variations of the different human. 

R.J. Talyor: I hear my wife talk a lot about how much she values brands that show all sorts of different ranges of models instead of always showing the, like, the stick-thin twenty-two-year-old wearing the product. Instead, there's a variety of different models that show actually how things fit. She's like, "That's the best. 

R.J. Talyor: I love that I can see that." 

R.J. Talyor: but to create that costs a lot of money for the brand. And so a lot of marketers have turned to AI models to generate different body types, different heights, different [00:28:00] coloring, racial differences, like all of th-those different things so that people like my wife can understand, "All right. 

R.J. Talyor: Ooh, that one would look good on me." But that hyper-personalization means that there might be mistakes in there and almost infinite options. Like, from a QA perspective, how does the human then have enough time to QA all those different variations? So I don't blame a brand like REI for not having looked at everything, but, you know, there is a chance, the risk, the brand risk of getting zinged on social media or other, there, so. 

Janice Hostager: has to take responsibility. 

R.J. Talyor: Yeah, right. Right. Yeah, it's tricky.  

Janice Hostager: Can somebody find out more about you or your company? 

R.J. Talyor: Y- yeah, of course. Well, we're Backstroke. You can just go to www.backstroke.com. And, you know, on the human side, I grew up swimming. I swam in high school, swam in college. I swam this morning. I still swim. And that's my form of exercise as well as mindfulness. And so when I went to start the company, I wanted to have something swimming related, and so it's [00:29:00] backstroke.com. 

Janice Hostager: Love 

R.J. Talyor: Yeah. 

Janice Hostager: No, 

R.J. Talyor: Yeah 

Janice Hostager: easy to remember. You attach a story to it. 

R.J. Talyor: Yeah. Be human. 

Janice Hostager: Yeah. 

R.J. Talyor: Yeah. 

Janice Hostager: All right. Well, thanks so much, RJ, for spending time with us, and, love what you had to say. A lot to think about too for our emails.  

R.J. Talyor: thank you so much. Appreciate it 

Janice Hostager: so here's a little challenge for you this week. Take a look at one of your pieces of marketing, maybe it's an email or even a social media post, and then ask yourself, "If I remove my logo, would somebody still know this came from me?" If the answer is no, don't panic. That's something that you can work on one piece at a time. 

Janice Hostager: So if you enjoyed today's episode, I'd love it if you'd subscribe or leave a review or share it with another business owner who's trying to stand out in a world that's starting to feel a little plain vanilla, right? Thanks so much for joining me today. For more information about anything we talked about, visit myweeklymarketing.com/167. Thanks for spending a [00:30:00] part of your day with me today. I'll see you next time. Bye for now