How To Navigate The Complexities of Generative Content and Copyright | Matt Ranta

Matt Ranta: change some colors, add a second checkout button above the, product not just down below where you know, whatever it might be. And go watch your analytics.

And did you increase things. Great. Keep it live. Did things stay flat? Great. Keep it live. Did you negatively impact things? Okay. Maybe go back and and take that out. Right. And see if that thing that you did was the actual cause of the negative impact. If. If you rebound or things stay low or whatever, maybe it wasn't right, like you can go lo-fi on your science and, and kind of like just, just test and iterate.

Matt Edmundson: Welcome to the e-Commerce podcast with me, your host, Matt Edmundson. The E-Commerce Podcast is a podcast all about helping you deliver e-commerce. Wow. And to help us do just that, I am chatting with today's guest, Matt Ranta, from Nimble Gravity, about how to navigate the complexities of generative content and copyrights. A bit of a hot topic right now. But before Matt and I dive into our conversation, Let me share with you a podcast pick, a previous episode that I think you're gonna enjoy around this whole topic.

Uh, very recently I chatted to Max Sinclair about the possibilities of generative AI. Do check out that episode. You can access our podcast pick and our entire podcast archive on your free, on our, not your free website, my free website, uh, ecommercepodcast.net. Plus, if you sign up to our newsletter, you'll have all this information winging its way to you.

Uh, the podcast pick, the notes, the links from today's show, they all get sent straight to you, to your inbox at no cost to you, which is pretty amazing. So if you haven't signed up to the newsletter yet, what are you waiting for? Head over to ecommercepodcast.net and join us now. Talking of Max Sinclair.

Max Sinclair was recently on the e-commerce cohort, which you would've heard me talk about if you're regular to the show. Basically, the e-commerce cohort is a, is something that we do. It's a, it's a monthly mastermind, a monthly membership, a monthly, it's just a monthly group where we all get together and think about e-commerce and uh, yes, it's called the e-commerce cohort.

And to help businesses like yourself deliver exceptional customer experiences. Um, We've got a course for you, a free course, a free resource that you can take. Doesn't take long, uh, called E-Commerce Cycles. It's a mini course. It walks you through my proven framework for building a successful e-commerce business.

I'm gonna show you the specific steps we take as a team in our own e-commerce companies so you can see exactly how to put these concepts into practice in your own business. And like I say, the good news is it is completely free, uh, and you can find out more information about that at ecommercecycles.com.

That's ecommercecycles.com, which is brought to you by our monthly mastermind e-commerce cohort, which of course is today's show sponsor. Now that's the show sponsor. Let's talk about our guest, Matt Ranta, uh, who is a digital dynamo, which I just, Matt, I dunno. I just, just sounds like an awesome title to me.

He's the head of practice at Nimble Gravity which revolutionizes digital transformation, e-commerce and strategy for clients ranging from startups to billion dollar giants. Oh, we've got the billion dollar giants in the Oh, yes. Uh, tackling dive diverse industries like healthcare, DTC clothing, and endangered species protection, which is just awesome.

Uh, Matt's expertise propels businesses into the future with data science, a great term, which those guys have coined and explained very well on their website. Uh, they do cutting edge e-commerce. Uh, his team at Nimble Gravity, they sort of redefined success by transforming the way organizations operate and excel.

So great bio that Matt. Great to have you on the show. Thanks for joining us, man. How are you doing?

Matt Ranta: I'm doing well. Thanks for having me on the show, Matt. I really appreciate it. Great name by the way, and looking forward to a fun conversation.

Matt Edmundson: It's funny you can't see my diary, but, um, we d if you've not figured it out yet, listener, we, we, we, uh, we batch record.

So we record like two episodes at a time just so we can, uh, stay on top of things. And our next guest is also called Matt, but he's just had to reschedule. And I was like, this is, this is the easiest day in the world for me. I don't have to remember anybody's name. I'm just gonna call everybody Matt, and it's awesome.

That's perfect. It's a beautiful name too. Absolutely beautiful name. You know what? This is not related to e-commerce, Matt, in any way, but. Last weekend, a friend of mine called Rich Rising was over from Dallas. I was celebrating quite a milestone birthday and bless him, he flew in from Dallas. Yes. And uh, we had a, a big, he was just, he was here for just like three or four days.

I mean a hell of a way to fly just for three or four days. And we happened to go watch an England football game and they were playing near where my mum lives. And so I didn't see my mum on my birthday. So I said to Rich, I said, let's just go to my mum's house and pop in and say hello. And he said, oh, that'd be great.

So we went over to my mum's house and Rich said to my mum, he asked her the question that I've never thought in all my years of living to ask my mum. And that question was this, why Matthew? Why did you name him Matthew? And I thought, I just assumed it's cuz it was like the first book of the New Testament Do you know what I mean?

It's like I, I assumed that that was probably, cause it definitely wasn't a family name, but I never thought to ask my mum that question. Anyway, it turns out when it wasn't that religious at all, it was all to do with an actor from the seventies called Matthew Hudson, which my mom took quite a shining to.

Matt Ranta: Oh, that's funny.

Matt Edmundson: So that's, uh, do you know why you are called Matt?

Matt Ranta: Uh, I, I don't, uh, exactly. I do know that, um, My mother and father made a deal where if I was a boy, my mom got to pick the name and if I was gonna be a girl, my dad was gonna get to pick the name. And I'm very happy to have born, been born a guy, uh, not for any gender-based reasons or anything like that really, but just I didn't want to have the name that my father was gonna pick out, which was Ursula.

Matt Edmundson: Okay. So, yeah, that, that's, um, Matt is definitely, is Matt calling, uh, every, all the Ursula's listening to the show gonna go, hang on a minute. Wait a second. Yeah, yeah.

I know. Some lovely Ursulas actually, especially in the digital industry. So Ursula's, if you're listening, uh, we do love you. Um, but Matt is a very cool name, so you should, you should ask your mum. Go, mum, listen, where did the name Matthew come from? And then just let me know the answer, cuz now I'm really intrigued why Matthews are called Matthew.

Um, and it maybe, it's, uh, it's worth finding out. So, Let's talk about generative ai, um, because it is a topic which is blowing up. It has been blowing up for quite a while. I say quite a while in digital terms. It's, we've been talking about it for more than a few weeks, so it feels like quite a while. Right?

Um, so generative ai, things like chat gpt, um, and so on and so forth. So how, what's your experience here, Matt? How did you get involved with all of this?

Matt Ranta: Yeah, same exact thing, right? Like it's been inescapable since probably December-ish or so, right? Yeah. Uh, but we as a, as a group have been involved with, um, you know, open AI's, uh, GPT models since GPT two Okay.

And have been actually trying to help folks create usable. Um, generative AI products from a standpoint of doing things like SEO work. Okay? Can you go in and have it create a, a huge number of SEO friendly content based products that surround your business? And at that point in time, it wasn't really capable of outputting much.

You know, we kind of got some funny results, to be quite honest with you. Mm-hmm. Uh, at this point in time, really, you can go in and say if you had a sister website, Right, and you wanted to rewrite content from your main primary site over to your second site where you were just trying to grab market share with a secondary brand.

But you're selling essentially the same things. You could go through and utilizing the APIs rather than just the, you know, graphic user interface. Yeah. Utilizing the APIs you could. Put massive amounts of content into the engine, have it rewritten and output in a safe way, uh, as long as you're using good prompts.

Right. Right. I think that that's a big key, is the, the prompt kind side of things. And we're seeing this topic explode not only from a written content standpoint, but. There are literally thousands and thousands of tools at this point, and they do everything from product photography to background imagery to ad campaigns, to, to you name it, right?

And so I think businesses, who aren't starting to capitalize on this are gonna quickly see themselves left behind because it's like nitrous for a car, right? Yeah. Like a business can start utilizing generative AI tools and accelerate their growth, accelerate their productivity, and do so unbelievably fast. And with high, high aptitude.

Matt Edmundson: Yeah. So, You mentioned this, I'm just writing notes, uh, Matt, hence the, hence the scribbling sounds in the background. Yeah. Yeah. Um, so you talk about using, um, AI to accelerate growth. How would you, I mean, just from a pure practical sense, if I'm, I'm an e-commerce entrepreneur, I'm listening to it, AI is something that is becoming more bigger and more and more in my face.

Right. I can't escape it. Yeah. So they we're, we're sitting there listening to you saying, well, you can use AI to accelerate growth. How, how could an e-commerce entrepreneur use AI to accelerate growth? What are some of the the key things that they could do, which is really gonna help them?

Matt Ranta: Well, sure. I'll give you a couple of, of more concrete examples in that, in that scenario, right?

Say, say you have to write blogs or articles, um mm-hmm. For, for your product, and this is something that from a content perspective, you should be doing anyway. That process, probably say you're a small business owner, right? That's an overwhelming process sometimes, right? Mm-hmm. I have to give up. Hours of time to be able to sit down and write an article or a blog and rewrite it and or I have to hire somebody and pay them, you know, a large wage in order to do that on my behalf.

And now what you instead need to do is be able to write a short prompt, uh, that describes how you want something written. Uh, imagine you are an expert, uh, in the fashion industry, and you're talking about men's active wear, uh, and you're describing, uh, the new trends in, um, you know, top layers, whatever, right?

Mm-hmm. And you go on and describe exactly what you want. Almost like a creative brief, right? And then you say, please generate a 1000 word article, uh, to this, and that will happen in seconds from a platform like chat Gpt or jarvis.ai or any of the other written generative content platforms that are out there.

My recommendation is that you don't just utilize that immediate output. You want to go back and edit it. You want to reprompt the engine in order to tweak and fix things for you. You might say, no, this tone is too serious. Can you please make it more lighthearted? You might go through as an individual and say, I would never use those words that turn or phrase, whatever it is.

Mm-hmm. And put it out and you've taken a process that takes hours down to taking minutes instead. Right. Yeah. Here's another one. Another one would be there are product photography platforms that are out there right now that allow you to take a very small number of photographs of a product. Could be a sporting product, could be something like a basketball, it could be a clothing product, like a jacket, whatever it might be.

Yeah. Then you can take those images and you can upload them into the AI platform. And instead of going out and doing a photo shoot, uh, with a bunch of models and a long timeframe, in order to get everybody set up, take the photos, go process them, put them back, go through and pick out which ones you want.

You tell the AI platform, Hey, put that jacket on a model of this ethnicity in this setting, and boom, you've got it right. Yeah. All of a sudden your jacket is on a, you know, an African American female in New York City hailing a cab. Yeah. And you didn't have to go do that photo shoot. You didn't have to pay for the photographer, pay for the model, anything like that.

You've saved thousands of dollars and hundreds of hours of time.

Matt Edmundson: It's interesting, isn't it that, um, uh, I mean, I was talking to, um, Literally just before we got into this call actually, um, I was talking to a friend of mine who's a copywriter. Um, she's bidding for some work and she's like, what do you think I should charge for this?

And we were just sort of going back and forth and, um, she's been out of, out of the workforce for a little bit cuz of kids. So she's now doing, you know, she's becoming back part-time. And we had this really interesting conversation because when she, when she, cuz she used to work for me, Beth, when she left, it was a case of.

Beth wrote everything from scratch, Do you know, what I mean, and would spend hours researching and thinking about blog posts and, and preparing arguments and thinking about sales funnels and the, you know, just the layout of the text and, and, and what's the, the route we want to take 'em down. And I was, I was showing her today how as long as she knows the prompts and can get that information, you know, that prompt written and, and teach chat Gpt Well, the, the thing is outputted inside of 30 seconds.

Yeah. And so, You know, her skill now is not necessarily just in copywriting, it's in prompt writing, um, and exactly getting that copy out and then editing it. Right. What, when you, um, here's a question for you, Matt. Um, we've got this thing now where, uh, I'm reading more and more where Google is, you go to say, chat gpt, you say, write me a blog post on men's activewear, or whatever it was you, you, you talked about.

So write me that blog post and. Um, Google is starting to recognize now whether this is AI generated and it feels like, um, it's starting to penalize AI generated content. Now, is that just a mere rumor? Is that actually true? And if so, how do we mitigate against that?

Matt Ranta: Well, yes. Uh, there are platforms out there, I'm sure Google's policing it.

There are things like GPT Zero, uh, that can be utilized in order to, you know, put any piece of content in and say, what's the likelihood that this was written with ai? And it will come back and tell you, you can go test it. Write, write, write something in chat GPT. Go over to GPT zero and see what it tells you.

Right? Then write something by hand. Go over and see what it tells you. And it's amazing that it can even do that. But yes, that is gonna happen and I think. I think that we as individuals are gonna have to get used to one. The fact that some portion of the content we have out there is probably generated by AI at some point in time, right?

But then we also are gonna have to get that mix right of how much should you come back in and edit that and change it and make it unique and, and touch it up because, I think the search engines might go, they might overindex on a tendency early on to penalize against that, right? Mm-hmm. Whereas businesses are going to probably overindex the other way and they need to not right?

Of like, I'm gonna have 'em do everything. I'm gonna have 'em write all my content. Right? Like recently, I think it was maybe even back in November, um, uh, uh, CNET got penalized and had a bunch of articles called out for having in, in factual un. You know, incorrect information in them. Right? And it was like 75 articles that they had had done as a test, uh, as to see is, is AI good enough to do this?

And they had to go back and, and recorrect those, right? That's a process problem that isn't necessarily an AI problem. That's, you never had sat somebody down, had them actually fact check what got put out, right? So there's probably gonna be some mix as we move forward in the future. And I think people need to.

Become probably a little bit more accepting of that because those tools really do allow us to accelerate and I think that it also, Would then get fed back into the AI engines and their intelligence will increase and it will become more unique. And I do think it goes to that prompt writing you were talking about, right?

That your, your former employee and, and colleague is now really good at. That's the challenge, right? Like are you coming in and asking a unique prompt? Are you telling the engine to write something unique that it hasn't written before? Or are you just asking a basic question like, is the sky blue. Yeah, give me 10 ideas for a 15 year old's birthday party.

Right? Yeah. Like, don't ask the basic question, augment it with a lot more information, and, and then you're gonna start to get much better content. Still fact check it, still rewrite it to be a little bit more like yourself, and then you're gonna have some kind of mix. It isn't that much different than you know, a copywriter and an editor at a magazine or a newspaper, right?

It's a source of intelligence, outputting a first draft, and it needs to be refined. Right. That's the way I would look at it. Yeah. From an SEO perspective. Yeah, same thing.

Matt Edmundson: It's really interesting because I, I, I was, uh, last week I had a, a, a a, well deserved, I would say, Matt week off, uh, week off work.

And I spent the entire week, um, either in my wood workshop or playing around with chat GPT prompts. So the two things I was sort of, and um, I've got to deliver some talks, um, from the stage coming up and I'm like, I wonder how, cuz normally. I would spend probably 8 to 16 hours prepping, say a 30 minute talk.

Yeah. You know, it's this, that kind of, uh, ratio. I'm like how can chat GPT? And there are obviously AI platforms out there. I mean, you mentioned Jarvis.ai, I think it's now called Jasper ai, cuz I think wasn't Oh, you're right. Yeah. Somebody was gonna sue them from Marvel, I'm sure for calling it uh, Jarvis. Yep. Um, and so, Uh, they now call it Jasper, which we used to use actually before chat, GPT four, which we predominantly use now.

And I was playing around on these prompts and it took my prep time down from, for the latest talk that I've written, what would normally take me eight to 12 hours. Took probably 20 minutes, um, yes, just going backwards and forwards. And it just outputted a series of data, which I can then go back and I can now edit and change, feed it back into the system and see what it it comes out with is.

It's incredible really. What it can do. So I guess one of the questions, and let's let's connect it to the title of the podcast. Yeah. Uh, I guess one of the questions in all of this is like, um, the same with the text, same with the image, right? So it's creating blog posts, it's creating product images, which are Max, max Sinclair's, um, stuff does at ecomtent.

And so, Am I, am I breaking any copyright laws by using what chat GPT has put out there? What's the what? What usage rights do I have, I suppose? Is this, is this something Yeah, if I, I guess one of the big questions, one of the big debates is if I say, um, add this product, my men's active wear to a. To a guy which looks a little bit like Brad Pitt Do you know what I mean?

It's like, and it, an AI goes and generates a sort of, you know, almost perfect version of, of Brad Pitt. It's all a bit gray, isn't it? Or are there some guidelines in this that we can follow?

Matt Ranta: There's definitively some guidelines that exist in the end user license agreements for the products. Right. So that's step one.

I think that this area, from a legal perspective, and I'm not a lawyer, nobody should take anything I say as legal advice. Right? Disclaimer. Yeah. But yeah, like from a, you know, from a legal area, this is gonna be muddy waters at best for a good while, right? Mm-hmm. So here's what I can tell you from like an end user license agreement standpoint, and it's kind of, you know, Thought provoking at a minimum may be scary for companies at, at, at, at a maximum.

So we did some research as a company back in February, right? So a lifetime ago in generative AI really at this point. But at that point in February, we, we researched and surveyed, uh, working professionals in the United States and, uh, elsewhere as well, but majority of them were in the United States and found that.

Of full-time working professionals in the US. 45% of the respondents who were familiar with generative ai, right, if they first had to pass a gating question of, are you familiar with any of these? Mm-hmm. Of of which around 50% of people were. Yeah. And then from there, 45% of those respondents were taking generative AI outputs and claiming them as their own.

Right. So they weren't saying, I did this with generative ai. Mm-hmm. They were saying, this is my work product. Wow. And okay. In that is a huge challenge. Right. So Jasper, uh, in their end user license agreement spells out very clearly, you may not utilize our content and claim it as being generated by a human.

So any of the folks that had done that from Jasper are creating a scenario where they can't copyright that work. They can't actually claim it as their own. If they wrote an article and put a byline on it as well, you know, Matt Ranta or Matt Edmundson, uh, wrong. You know, it's, it's not theirs. It doesn't belong to them.

If you get into open AI end user license agreement, You'll see some even more interesting things. Kind of where they go is if you output content that is a hundred percent unique. Uh, then you can, then you can start to utilize it. Mm-hmm. Um, but if you and 50 other people ask it the same question and get the same output, that output is actually public domain.

It is not individual, and it cannot be copywriting claimed as as your own. So that's where that prompt writing actually becomes even more specialized. Right. And more critical. Right. If it's unique, prompt, if it's completely new and, and of, you know, Brand new thought process, then that content can start to look and become yours.

Hmm, I haven't researched the end user license agreement of, you know, thousands of these ais, but here's what I, here's what I would recommend to people. You've gotta go read it. You've gotta understand, can I actually utilize the output of this as my own? Some of them are very good about calling it out, right?

Like if you're buying a logo or something like that, that AI has generated, they're very good about saying you have, you know, Unrestricted use and perpetuity for this. You're buying it as as your brand mark, right? Yeah, go ahead and do it. Others are burying it in very, very deep places, right? So I would recommend go read it, go have somebody on your legal team if you're large enough to have one of those.

Read it and understand that deeply. The second thing I would do from just if I were a company, Running a business where any generative AI could touch it is I would have a policy for my employees. Right. And I personally wouldn't make this policy restrictive in the sense of, Hey, you can't use generative ai.

Mm-hmm. What I would say is you can use generative ai, but we have to know. Right, right. Like we as a company have to be aware of what's going on. You can't claim a work product that AI built as your own and then Right, have us going out and copywriting it, et cetera, et cetera. Right? You're creating challenges for that organization, so you need to have something that acknowledges that AI is here.

It's not going away. People are going to use it. People want to be more productive, but you've gotta have some level of transparency within your organization that allows you then to make decisions downstream from where that content was generated and how it's being used, that people are fully aware and not. You know, setting themselves up for a surprise six or eight months or a year from now.

Matt Edmundson: Yeah, that's really interesting. That's really interesting. So is that what you guys do at Nimble Gravity? You've got a sort of a policy on, on generative AI usage that your staff follow?

Matt Ranta: Uh, we don't yet, quite frankly, and it's something that we're talking about. Um, and it, it. It's something that we utilize, um, fairly regularly too. Mm-hmm. Right. Like that research process that you're talking about where you're, where you're going in, uh, and doing, you know, prep for a speech or prep for understanding what are the differences between software platforms or anything like that.

You can do that in chat, GPT, right? Mm-hmm. Uh, you need to go fact check that. Can you help put together a presentation? You bet you can. Can you put together a blog post? Sure. You can. Can you put together materials that go inside of, uh, a, a client facing presentation? You, you sure could, right? Mm-hmm. But how much do you want to edit that?

And so I, I think that we, we need to practice what we're preaching for one, uh, and, and two, um, it's ever evolving. And so that's one of those things like a privacy policy. But you're gonna have to go back and revisit, right? Like so, okay. Gdpr. Okay. Brexit, no UK's own thing. Uh, yeah. California's, CPRA, uh, Colorado that I live in now all of a sudden has a policy, right?

Like, and they just keep piling on each other and. You're gonna have to figure this out as a business.

Matt Edmundson: Yeah, it's crazy. Uh, part of me, I'm, I'm sort of sitting here wondering, I wonder if I said to chat GPT write me a policy whether it would actually create the policy. It probably would, uh, if I put the right prompts in.

Um, one of the things I found actually with Chat GPT, um, and again, it, I guess it's the same with the other platforms, is if I give it copy. Of, um, copies of, or chunks of content that I have previously written in my sort of style, my tone of voice. Yeah, it can mimic that quite well. When it's output in the content, it can sound quite, quite like me.

Um, which minimizes my editing, which I'm assuming will make it a bit more unique cuz it's, it's sort of. My Do, you know what I mean? My tone of voice. And I can, I can exactly throw that in there, which is, which is quite handy. And it's quite nice because you can do that with brand as well, can't you? You can throw brand voice in there and brand values and it tends to output it in that way.

Um, which i, I quite like. So what are some of the, the creative, I mean you guys are using it a nimble gravity? I'm, I'm, do you use it on behalf of your clients? Uh, what are some of the surprising ways you've seen, I guess, uh, uh, generative AI being used? Yeah,

Matt Ranta: so you know, the example that I called out previously of, you know, rewriting, um, SEO friendly content from, uh, one site to another, uh, is something that, that we've, uh, worked with and, and created models around.

Uh, and I would say that written generative AI content is probably the largest area. Right now. Mm-hmm. Right. Uh, some of the, uh, photography based kind of stuff is still a, a little bit new. Mm-hmm. We, we don't do graphic design or anything like that, so we're not necessarily looking at some of those, um, you know, brand or campaign level kind of AI that allow you to create a, a whole bunch of templates.

Um, but it seems like written content is, is the largest thing that's happening right now. Yeah. It's for real applicable use. Yeah, yeah, yeah. And that goes to, that goes to code too, Matt. So I just wanna throw that in there, like actual programming code that goes live in websites. So

Matt Edmundson: Yeah, I, I know one or two of our developers, they've started to use chat gpt to generate code for them, which is quite fun, quite funny.

Yeah. And what what's fascinating, I mean obviously they still need to understand the code. My, uh, my friend who was a copywriter still needs to understand is this copy right? Is it good? It still needs that human interaction. Right? Um, but the interesting thing, cuz when I was talking to my, talking to Beth, a copywriter friend, at first you kind of go, well chat GPT's gonna take over my

job, it's people aren't gonna need me. And I'm like, well, not, no, not yet. Not quite. Um, and actually what I think you can do is probably put your fees up quite a bit, um, for what you were charging by throwing the fact in there that you're gonna use chat GPT and actually you're now prompt engineering to get the best out of it.

Right. Right. Um, because that's a highest skillset, I think, because there are fewer and fewer people who, who know how to write the prompts. It seems for AI to get the stuff out of it. And so, yeah, it's a, it's a fascinating one, this whole world of where do you think it's gonna go to? Because like you've now got Elon Musk coming out and saying, oh, we need to stop everything and throw, slam the brakes on.

Um, where do you think it's all gonna go to?

Matt Ranta: Yeah. Yeah. It's interesting, right? He, he wants to write his own thing and he's also promoting, putting the breaks on everybody else's. Interesting. Um,

Matt Edmundson: Elon Musk for president. Yeah, sure.

Matt Ranta: Yeah. Um, You know, where is it gonna go? It's ra, it's so rapidly advancing, right? Mm-hmm. I do think, I do think prompt engineering is gonna be an actual skill. Mm-hmm. Probably an actual job, right? Yeah. And. You're not necessarily gonna be able to be a good, prompt engineer for production code, as well as be a good, prompt engineer for written, readable content. Mm-hmm. As well as like, right.

Like this is almost like you're the director of a movie. Yeah. And you're trying to tell the entire crew what you want out of it. And some people are great at directing westerns and other people mm-hmm. Are great at directing sci-fi. And some people do really well with documentaries. Right. But you can't take a documentary director and say, all right, we're gonna put you in this tent pole, blockbuster, uh, you're directing the fifth Avengers film.

Go like they're going to, they're gonna fail at that. Right. So becoming a prompt engineer is gonna be. The job, and I think your education is gonna have to shift in order to do that, right? Like this is causing unbelievable upheaval in school boards and all kinds of stuff too. Oh my gosh, my students are cheating.

What's going on? Like, how is this gonna happen? Stop that. Like stop it now, and start teaching them how to be prompt engineers and start teaching them how to be subject matter experts. Mm-hmm. Because regardless of of whatever happens, you are gonna have to be a subject matter expert to be. Good at being a prompt engineer.

Yeah. Can you use AI to help you become a subject matter expert? You bet you can, right? Mm-hmm. Like you can go do research like we were talking about, but again, it goes down to prompts. Yeah. So teach people how to use it versus being afraid of it and thinking it's gonna steal and plagiarize everything.

Right, right. Like teach them how to be good, curious, inquisitive humans that find a passion around something and then turn that passion into becoming. Somebody who can write fantastic prompts and direct these tools to your advantage. Yeah. So that's what I would say.

Matt Edmundson: That's such a powerful point, Matt, and I think it's very well said.

And, um, I'm smiling because I remember, you know, when I was back at school, uh, and they just invented electricity, um, they were, they were, they were like, well, you know, you, you, you've got calculators now. We had scientific calculators and I remember when the first calculators came out that could draw graphs.

Yeah. And they're like, well, you can't take those into the exam cuz that's, you have to figure out how to draw the graph for yourself. And it's, I think education when it, when it realizes this is where technology is, how do we help our kids utilize the technology to be smarter rather than try and stop them?

Um, right. Because they're kids. I mean, you know, adults are using it all the time to make their life easier. And we expect our kids to go, well, I can't do that. That's just not right for my exams. It's just this ludicrous, isn't it really? In a lot of ways. Yeah. So, uh, it's fascinating. Fascinating. So you've got, um, Let's just start.

I'm curious, uh, in the, in the last few minutes of this show, uh, Matt, yeah. If I can dig into, I did look at your, um, website, which is nimblegravity.com, and I was really curious by this phrase, data science, which I, it is quite a generic phrase, but you use it in quite a specific way in the sort of the, the junction of the three circles. Um, mm-hmm. What, what do they call that graph?

Matt Ranta: Uh, a Venn diagram.

Matt Edmundson: Yeah. This is where I, I should have paid more attention in Math rather than looking at the, the, the, the sine wave on the calculator, uh, the Venn diagram. So just explain that and how did you come up with this?

Matt Ranta: Yeah. So you know, it. We talk about pragmatic data science, right?

And that's, you know, understanding what you're trying to do From a, a general business hypothesis, what can you do to test quickly, to be iterative, to, you know, extract something in a very short period of time to then, Be intelligent about that and move on from there, from whatever you learned and create the next phase and, and you know, kind of creating this iterative cycle.

Yeah. Just like you would in an agile programming environment. Right. But doing it with data science tests. And projects and moving quickly and, and forward and, uh, iteratively. And, you know, the venn diagram that you're referring to on, on the site kind of shows, you know, artificial intelligence, machine learning, math, and one circle, it talks about technology in another.

And then, you know, brings in business and industry acumen into the other. And, you know, that's subject matter expertise, right? Mm-hmm. And then that center overlapping area is, Where data science actually happens, right? So you're, you're taking the knowledge of the folks within the industry, within the business and their, their specialized subject matter expertise over years.

Right? And you're combining that with technologies that allow you to move and analyze huge, large quantities of data. And then you're applying, you know, uh, Mathematics, algorithms and equations with machine learning and AI on top of that, and, bringing it all together into, into the world of data science.

And then you're able to output insights from an organization through their data that they might not have really seen before, right? Mm-hmm. Like. They might have had gut feelings that, you know, that the long-term employees that that really understood the business in, in total might have been expressing these things.

But you can actually put out the goals and the percentages and, and these kinds of things you can do. Creation of propensity models, how likely is a customer gonna be to buy from us? How likely are they gonna be able to buy a second time? You can output, you know, actual, um, modeling around, you know, people who interact with your business and footfalls and what's going on there.

Any number of things really. Um, and. It's, it's a lot of fun. You can take alternative data sets and understand, um, you know, what are, what are people doing in their day, right? Mm-hmm. From foot traffic and stores and, you know, where else are they going? Um, that we might be able to, you know, modify our marketing messaging and placement in order to capture these people.

And you can unlock these kinds of super interesting approaches to interacting with your customers that you probably wouldn't have done through traditional tools. Mm-hmm.

Matt Edmundson: Yeah, that's really interesting, isn't it? And I, I, I, I get sucked into it, Matt. I do. Because it, you just get drawn into all the, all the, all the possibilities that this sort of technology and thinking now brings to you really, and the, and the possibilities that are out there.

Yeah. I guess one thing, um, that sort of springs to mind in all this is, It sounds expensive and complex, so if I'm just starting out in e-commerce. Um, I'm curious to know your thoughts on this, cause obviously you're dealing with companies doing billion dollar turnovers and so it's kind of like mm-hmm.

If you are starting out and you're sitting there and you're listening to people going, well, we can, you know, do this data and all, and all these sort of propensity models and all that sort of wonderful stuff, but you're like, dude, I've got like a thousand people coming to my website a year at the moment of which 15 are buying and I'm trying to build the business, I think.

It, it feels, I think it can feel quite overwhelming for a startup Do you know what I mean? An entrepreneur to think, oh, I've gotta have this, I've gotta have this, I've gotta have this, and, and, and all these tools that the big guys have got, how am I even gonna compete? And I wonder what your thoughts were on that, because it, I mean, in some respects, it's not actually a million miles away from the guy that's starting up right now, but I, I can understand that it, it feels like a massive world out there.

And so I just wondered what your, your thoughts were on that for the little guy.

Matt Ranta: Yeah. So a couple things that I'll say around that. There's a lot of platforms that have been, that are data, data science platforms that have been productized, right? That allow you to do things like, um, price testing and creating pricing models, right?

And so as a smaller guy or gal running a business, right? Um, it you don't necessarily have to, um, Have a huge data science budget and hire a bunch of people and, you know, have this massive team, or go to an agency and have them do that on your behalf, right? You can utilize some of these, these tools that are out there.

Uh, and you know, like the pricing model that I'm talking about, um, there's a platform called Intelligence, uh, that, you know, we've been talking with and, uh, they do this right? And, um, there, there's some interesting things out there like that. But here's the other one, right? Like being a data scientist, I would argue, right?

Like I think we think of it as folks who are, they have math degrees, they have computer science degrees, they're working, you know, with huge data sets and environments like Snowflake or whatever, and Right. Like those people are professionally labeled as data scientists. Mm-hmm. A true data scientist is somebody who is curious, who's willing to test and experiment and try things and create a hypothesis and go out and execute against that and see what the results were, and then come back and iterate on that.

Right? Yeah. And you know where you can do that Excel. And Google Sheets, right? Like on the low, low end of things, uh, you can go back and do regression studies in Excel, and there are a thousand tutorials on the internet about how to do this. It's built into Excel. You can understand what happened with your data in the past through a regression test, and you can even compare this to.

Data sets that are freely available from, say, the government, right? Mm-hmm. Uh, I did this back in, oh gosh, I wanna say it was 2008, with, uh, a couple super smart business analysts. They were data scientists, but at that point in time they were called business analysts. Um, and, you know, we did the, the bulk of it, um, in, in Excel and.

What we found at, at this point in time is we were doing a regression analysis of appliance sales, uh, and comparing that to housing starts, uh, which are available government metric here in the United States that you can download for free off of, uh, off of government websites. And we were looking to understand, um, how the correlation between housing starts in the areas we, we looked at.

Met up with appliance sales. Right, right. Like that makes sense. You're building a new house, you gotta build, you know, you gotta buy appliances. What's the gap? Right. Like that's really what we wanted to understand and, and know when we should really be going to market and talking to contractors and, and trying to mm-hmm.

To land large sales deals. And that's a hundred percent doable in Excel. Right. Wow. And I think it's more. Go figure out what your hypothesis is. Mm-hmm. And then figure out the environment that you ha, you know, kind of need to utilize in order to do that and get yourself, you know, somebody curious and, and smart and good with manipulating data in multiple types of platforms and places. Mm-hmm.

Matt Edmundson: That's top advice. So let's start small. Just start doing it, I think, um, totally. Because e-commerce is both art and science, right? It's not, and you, the, the sooner you can get your head around understanding some of these things, the better, I think. Uh, yeah. And, uh, and, and if Excel can help you, and Google Sheets can help you have at it. I mean, I've spent many an hour in a Google spreadsheet.

That should almost be a t-shirt slogan. Um, listen, uh, Matt, uh, I'm aware of time, man, and, and I feel like we're just starting to scratch the surface. So let me ask you one final question, if I may. Yeah. So you're a data scientist. You've got this company which is helping, you know, all manner of different size organizations.

You've seen a lot, you've experienced a lot. What's the one opportunity you think that e-commerce entrepreneurs are not taking advantage of on the whole, that they probably should be doing right now to have the biggest impact on their business? From your point of view, from a data science point of view, what, what do you, what do you see in all of this?

Matt Ranta: Oh, wow. Um, That's, that's a great question and obviously a big challenge, and I think it's kind of one of those overwhelming things for, for e-commerce folks, but I'm gonna go back to a, a tried and true standard that I unfortunately don't see enough people executing well enough on a consistent basis. And that's AB testing.

Right. Okay. And this is a surefire way to increase the outputs of your e-commerce operation, right? Mm-hmm. Uh, yes. Sometimes that requires a, a good amount of traffic, uh, and, and you need, you know, a certain number of, of visits, uh, per year to utilize platforms that are gonna do that. And I think that that's where people get a little bit scared too, from an AB testing perspective.

Mm. So if you can utilize a platform, if your website has enough traffic, uh, in order to be able to, to really get impact, uh, from an AB testing platform, uh, a hundred percent you should be doing it. You should have a team around it. You should be working iteratively. You should be, you know, studying the results and pushing live new features and functionality and layouts and user flows.

All year long, like do as many tests as you can. If your traffic isn't large enough to support a platform like, you know, an optimizly convert.com, whatever it might be, uh, I'm gonna go lofi again, right? And just be like, be brave. Start track what you're doing, like. You know, write it down somewhere, put it in Jira, put it back into Excel, or a Google sheet of mm-hmm.

You know, this is my theory, here's my test. I launched it on this. And go watch your Google Analytics. Yeah. Or your free analytics platform. Right. Change, change some colors, add a second checkout button above the, you know, the, the product not just down below where you know, whatever it might be. And go watch your analytics.

And did you increase things. Great. Keep it live. Did things stay flat? Great. Keep it live. Did you negatively impact things? Okay. Maybe go back and and take that out. Right. And see if that thing that you did was the actual cause of the negative impact. If. If you rebound or things stay low or whatever, maybe it wasn't right, like you can go lo-fi on your science and, and kind of like just, just test and iterate.

And in that scenario, I wouldn't be doing a whole bunch of tests, but I'd be doing some tests and I'd be being brave about it.

Matt Edmundson: Yeah. No, it's very good, very good. I, I love that, especially your product page. Um, I don't feel like people product test, uh, split test their product page enough and you can run different ads to different versions of your product page.

You know, you can, there's all kinds of ways to play around with that and, and definitely do that and get involved with it. Um, Uh, I like that. I like the simplicity of that. Matt, I'm not gonna lie, I was kind of expecting you to say something in ai, but actually going back to split testing makes an awful lot of sense.

And then as you were talking, I'm going, that's, that's actually the right answer. Uh, I I thought that was, uh, very well said, sir. Very well said. Thank you. So, uh, no, listen, um, Matt, listen. How do people reach you? How do they get a hold of you? If they want to connect with you guys, ask you some more questions, what's the best way to do that?

Matt Ranta: Yeah, you, you already mentioned the website, nimblegravity.com, so that's definitely a great way to, to get ahold of us. The other easiest place right, is LinkedIn. Uh, feel free to to reach out to the company or follow us there or reach out to me directly. Uh, I'm on there, Matt Ranta. Uh, and happy to connect with folks and chat with them about any of their e-commerce, digital AI challenges, whatever it might be.

Matt Edmundson: Fantastic. Absolutely do go connect with Matt on LinkedIn. I need to check, see if we are LinkedIn, LinkedIn Linked, linked on LinkedIn. I know it's not easy to say connected on LinkedIn. Um, uh, we we should definitely do that. Listen, we will of course link to Matt's info in the show notes, which you can get along for free along with the transcript at ecommercepodcast.net.

Or it will be winging its way into your inbox. If you've signed up for a newsletter, Matt, listen, genuinely really appreciated the conversation. I'm loving this stuff right now. Um, I, I said at the start, the show sponsored by the e-commerce cohort, which it is. And in the cohort, which is our sort of monthly mastermind group, uh, we've just done a whole thing.

I run, I, I do this thing on, on chat GPT, on how to start an e-commerce business. How to use chat GPT to come up with the name of the company, the products that we're gonna sell, what's gonna differentiate it, how we do customer research, how we write product copy. We had a, we had a bit of a laugh actually with that.

Nice. And so, Yeah, if this is something which has perked your interest, dear listener, do check out ecommercecohort.com. There is like a, a sort of a freezed trace. It's like a dollar trial, I think, or something like that. Check it out, uh, on the, on the, the chat thing that we do. It was great fun actually, but Matt, genuinely appreciate it.

Love this stuff, love learning lots of stuff about ai. I think it's a really interesting thing. So thank you so much for joining me, man.

Matt Ranta: Thanks for having me. Great conversation. Really enjoyed it.

Matt Edmundson: No, it's been great. Great. So thanks again to Matt and also, again, a big shout out to today's show sponsor, the e-commerce cohort.

Uh, the free training is at ecommercecycles.com that I mentioned. And if you want to check out more about the, uh, chat GPT training that we did. That's at ecommercecohort.com. A lot of domains, but they're all linked on the website. Now, be sure to follow the e-commerce podcast wherever you get your podcast from because we've got yet more great conversations lined up, and we don't want you to miss any of them.

And before we wrap up today's episode, let me take a little moment to invite you to become a part of the show now. If you're an e-commerce entrepreneur or an expert and would like to share your insights with our audience, we would love to hear from you. Or maybe you just know someone who would make a great guest.

Do send 'em our way. Just head over to the website, ecommercepodcast.net and get in touch with us. All the information is on there, and in case no one has told you yet today, you are awesome. Yes, you are. Created awesome. It's just a burden you have to bear. I have to bear it. The other Matt has to bear it.

And just about every Matt on the planet has to bear it. And if your name's not Matt, you've also gotta bear it as well. Now, the e-Commerce podcast is produced by Aurion Media. You can find our entire archive of episodes on your favorite podcast app. The team that makes this show possible is Sadaf Beynon, Estella Robin and Tanya Hutsuliak.

Our theme song was written by Josh Edmundson, and as I mentioned, if you would like to read the transcript or show notes, head over to the website, ecommercepodcast.net, uh, where you can also sign up for our weekly newsletter and get all of this good stuff direct your inbox. That's it from me. That's it from Matt. Thank you so much for joining us. Have a fantastic week wherever you are in the world. I'll see you next time. Bye for now.