It’s a simple concept really – people want to buy things that other people have already bought and have given their approval. Good recommendations for a product means more people will buy, and bad reviews mean the product might flop. But here’s the thing: As an ecommerce business, you almost certainly need to have product recommendations available because of these stats:
So as an ecommerce company, how do we leverage product recommendations to maximize conversions and order size?
Today we welcome Michael Prichard, CEO of Skafos.ai, to tell us about a new way to do product recommendations.
Izach Porter: All right, you’re listening to the Deal Closers podcast, brought to you by websiteclosers.com. A show about how to build your e-commerce business to be profitable, scalable, and one day even sellable. I’m Izach Porter and on the show today, the CEO of Skafos.ai, Michael Prichard joins us to talk about product recommendations on your e-commerce website. It’s really a simple concept. People want to buy things that other people have already purchased, and that they like. It’s like word of mouth marketing but on the internet. Good recommendations for a product mean more people will buy, and a bad review means the product might flop. But here’s the thing as an e-commerce business, you almost certainly need to have product recommendations available. Because of these following statistics. 54% of retailers reported product recommendations as the key driver of the average order value in the customer purchase. 73% of customers are more likely to buy based on personalized recommendations, and 37% of shoppers that clicked on a recommendation during their first visit returned compared with just 19% of shoppers that didn’t click on a recommendation. So as an e-commerce company, how do we leverage product recommendations to maximize conversions and order size? Let’s find out from Michael Prichard. Hey, Michael, how’s it going today?
Michael Prichard: It’s going fantastic Izach. Thank you for having me.
Izach Porter: Yeah, man, thanks for being on the show.
Michael Prichard: Absolutely.
Izach Porter: Really glad to have you. So look, you know, I speak with a lot of e-commerce business owners that are thinking about how to sell their business getting ready to sell their business. And this concept of increasing conversion rate increasing AOV, increasing customer lifetime value. It’s like top of mind for all of these companies. And the feedback I hear pretty regularly is that the tools that are available, kind of Shopify, generic tools just aren’t really cutting it. And so when I when I learned about scaffolds, I was really excited because it sounds like you’re doing some things just very different and much more intelligently than how it’s been done in the past. He’s give us a little overview of what your company does and how it works?
Michael Prichard: Yeah, absolutely. Thank you. So we are on a mission to help shoppers find the products they want, in a more delightful way. And we do it through simply by asking the customer if they like something or if they don’t like something very much like what you’d find in other social media type of paradigms on say, Facebook and whatnot or even Tinder. We’ve been called the tinder of shopping, which is an internal joke, which we’ll talk about another time. Here, swipe right, swipe left kind of thing. But the end of the day, what we’re solving is not only are we personalizing the shopping experience for customers using traditional recommendation kind of algorithms, we’re also using interactions from them to really understand context as to why they’re there. The one of the challenges, just gonna jump ahead a little bit here is that while product recommendations work, especially on say past products that have maybe been sold, or look alike, kind of audiences and things like that, as you know, privacy is becoming a much bigger topic in the world in general and with online commerce. People are individuals, people want agency, they want to control their journeys, etc. With all of that, it’s much harder for traditional recommendation systems don’t work. And our special sauce is that we basically use behavior and interactions to quickly surface up products that that someone may want, which is actually a very simple concept that no one else is doing except for us and Amazon actually
Izach Porter: You are kidding, really interesting. So consumers or you said it’s interacts kind of like a social media sites comes to mind thumbs up, thumbs down smiley face emoji’s, that kind of that kind of?
Michael Prichard: Yeah, I mean, thumbs up, thumbs down or more like this, less like this. Something has been working really well. Is it within the PLP, PLP is the listing page. So for the Shopify folks on the phone, Shopify folks on the podcast I guess. You know, the product listing page like adding a little see similar button that in real time resorts the page so to speak, or you can drop in a more like this or less like this or not my style within a rack slot on the PDP for example, very, very simple approach of kind of making the experience more dynamic is what we do.
Izach Porter: That’s awesome. Yeah, I think as competition increases, empowering founders and owners of these companies to be smarter and get better interactions with their customers is, it’s good for the customers, which supports the brand, which drives the value for the owner. So hey, so going back to some of the statistics I mentioned kind of at the beginning, driving average order value and customer purchase, more likely to buy more likely to return. And what statistics are you seeing that kind of matter in the space? What do you focus on, as you think about how scuffles is interacting with your customers?
Michael Prichard: Right. So, I can only speak from what we’re seeing in the market. I mean, you can find stats in general, obviously, you quoted a few. I mean, another one is that 35% of Amazon’s revenue comes from personalized a recommendation actually, which is pretty huge. You know, what we’re finding right now is a couple things. One, when people interact, or when people are given agency at least asked, Hey, am I be in the merchant on the right track? Right. That conversion rates will jump up by 100%. So yeah, it’s pretty wild. So we had one customer that deployed an earlier version of one of our solutions called product finder, and they jumped 105%, the jewelry company, actually out of Australia was like, my favorite place right now. Because we seem to be very popular, they’re world famous in Australia by the way, and 105% increase of CVR. And they did I think, 85% on revenue per visit. So basically, more folks bought, when they visited versus not using our platform, etc. One thing we did notice, actually, which you talked about AOV, which is the average order value, right? So there’s average order value is lifetime value. All those are very important. We did notice on the problem, very open everyone, we’re figuring some of this stuff out, as we’re doing this, that even though we have higher conversion rates, and revenue per visit, average order value sometimes went down. And we’ve been trying to figure out what was happening there, and either we’re getting people to products much faster with originally, or maybe we need to expose some of the higher end sort of products within the recommendations, which is something we’re working on right now. So, but when you cut it back, and you look at it, from a point of view, like I’m converting more, I’m driving more revenue per visit, the lifetime value of that customer is going to be higher. So maybe AOV is not quite always the thing I want to be paying attention to is, it’s like, are they coming back? And, are they buying?
Izach Porter: Yeah, yeah, that’s kind of the question I was I was getting to I think that’s a great point is that what is this statistic that matters? And are you able to drive that sounds like, LTV, I think for me is critically important, because a returning customer, you don’t have to pay to acquire twice typically?
Michael Prichard: Yeah, that’s something we’re actually working on with another customer of ours called , they are a leather company. So they supply if you ever bought a couch, a leather couch from Pottery Barn, or Restoration Hardware or written a jet, which I have not, you know, wood leather seats, and most likely came from, from them. They also have an e-commerce site that does, bags, various accessory products that go with that. And we’ve noticed that, if we can get someone off to the first sale, obviously big, but the first, a second sale is really hard. If we get them to a second sale third, and fourth, sales are much easier. So like, so how do we do that? And that’s by quickly understanding context, understanding why the person’s there, what they truly like, we may not know exactly why they like it, we just don’t like certain shapes or textures or whatever. And then being able to kind of suggest to them products that either may go along really well with what they’re doing or maybe, we try it, we try to capture them in a moment so to speak. And, and I think that’s where we’re gonna see a lot of huge wins is because we can, we can understand a lot better than traditional personalization systems, which in my opinion, are very generalized and very static. And honestly, past behavior doesn’t dictate future behavior. You know, people are, so yeah.
Izach Porter: So the scuffles weighted similarity algorithm. How does that kind of fit into making, bringing all these concepts to life?
Michael Prichard: Yeah, that’s top secret. I can’t, I’m kidding. So we can’t talk. We gotta bleep that out.
Izach Porter: I want you to spill the beans on our show.
Michael Prichard: I mean, it’s so like we’re, so basically what we do is, you know, here’s the beautiful part about is that we don’t need to know a lot of data like about your customer, right? We don’t need to have a lot of history. Now, if you have history, like past order data and things like that. Great, we can take that into our system, but all we need is a catalog. And once we get the catalog, which is really, not hard thing to do, we have a series of machine learning models, and I’m gonna say, AI models, which we can get into, that we train. So we train a variety of say, object detection models on the images, right. We train a series of similarity models on the images on the metadata which includes like, the description, price, the titles, things like that. We train off of, like if we get past order data, things like that. So we basically create a series of models and our special sauce is, the way we kind of string those models together in a weighted system. So as you start to interact with our platform, our almost AI for lack of a better description, and those data scientists can yell at me later, but like our AI will start to automatically wait different pieces of the of those models based on how people are interacting, right. So like, don’t get too technical. But we’re not, I wouldn’t even consider the recommendation system as more of a search system. Like we, because when you make a request to our platform, it’s really making a search request. It’s like, Hey, here’s the input of the search. Give me back what I’m asking for. And then based on this, as many times as you start to search with us, we just refine the results. Does that make sense? That’s really how we work now.
Izach Porter: The search keeps getting smarter and smarter the information you are sharing.
Michael Prichard: Yeah, we’re sorting in Research Company that basically happens to bring it up in an interactive UX, which is like, a like and dislike, and things like that. So that’s what makes us unique.
Izach Porter: Got it. So I think we’ve all heard about companies that manipulate reviews. Is that a problem that you see across the industry? And, how does that kind of play into what you’re working on?
Michael Prichard: Because like we do, we do integrate out of the box with right now with the OPPO. And with Luke’s and we’re about to really stamped. And we’ve started to add as part of our way to similarity algorithm. But it’s not that big of an effect right now. So I mean, to be honest with you, I don’t I think reviews are hugely important. I think you should definitely have them on your site. People do use them. I’m, you know, we’re all consumers. We’re all shoppers, right. So we do use them, I use them. So I think you definitely need to have them, I would make sure that you don’t fake them, like do not like, if you get the call from like, for example, we have a Shopify app in the App Store. And we get maybe, I don’t know, let’s just say monthly get hit up by somebody, like I have a review service, I’ll get you 5000 reviews for five bucks or whatever. And like, that is the worst that you could do. Because people know if it’s real or not. So I would I would definitely work on a post purchase series or workflow that encourages, like Izach to give me a review, right. And want real review. So I think if you can get real reviews, and don’t get caught in the trap of getting fake ones, which some people do, let’s admit it, they’re going to work very well. I guess the short answer is like, people aren’t dumb.
Izach Porter: Yeah, right. We’ve all read reviews where you’re like, that’s not
Michael Prichard: Yeah, I’m not convinced that’s like a, that’s someone got paid. I mean, so it’s a review. Therefore, I actually want to talk about one thing, actually, I’m not sure the next question is, I apologize. But back to like data we’ve seen, which I think is really interesting is that I’m not sure if we’re like taught this, like getting no, like the word no, or, people seem to, like have an adverse reaction to like, oh, I don’t want my customer to tell me no, I don’t want to hear no, right. But what’s interesting is in our data, especially with our product finder solution, we found that when someone says no, like, not my style, or don’t show me this, that the conversion rate jumped 3X like one customer we were hitting 2% to 3% on just regular conversion rate. When people said no, and they removed that those items from the sorting and the search, and in that particular recommendation, it jumped to 6% like it was crazy. We’ve seen it happen over and over and over. And so we start thinking about, why is that happening? And so, we have actually a blog post on this called “The power of No”, and the theory we have is twofold. One, have you ever read Chris Voss’s book “Never split the difference”? Which I recommend…
Izach Porter: No, I haven’t
Michael Prichard: I recommend everybody read.
Izach Porter: I’ll check it out.
Michael Prichard: Yeah, it’s fantastic. I hate reading and it’s gonna sound horrible to people. I do read obviously, or maybe not.
Izach Porter: You got Chris Voss book?
Michael Prichard: I read Chris Voss’s book, because it was great. It was a good book. It wasn’t like that typical business book that’s like, you can read it in 20 pages and then like 380 pages of crap like you’re like, Okay, this is not you’re wasting my time right now. His book actually is legit. So basically, he’s an ex FBI negotiator. So this guy was negotiating people’s lives, right. And they would purposely ask questions to get the person on the other side of the table to say no. Why? Because it gave them a perceived sort of a perception of control, right. So all of a sudden, if I say, I don’t want to see this particular pants, for example, is psychologically set me up to have control, which is what people want, people want agency. I want agency, right.
Izach Porter: Interesting. So, I say I don’t want to see these pants. All the pants I see, after that I’m like, Oh, these are the pants that don’t want to see.
Michael Prichard: Like, I don’t want to see that.
Izach Porter: I said, I didn’t want to see the blue pants.
Michael Prichard: Yeah.
Izach Porter: I want black pants.
Michael Prichard: Yeah, yeah, you’re liking them with the polka dots. So you know, like, I don’t want to see the ones that polka dots kind of thing. So that’s one theory. And the other one I think is happening, especially in larger catalogs, is what’s Hicks law, which is I’m gonna paraphrase it the best I can. It’s basically, the more choices you give someone, the longer it takes for them to make a decision kind of thing. And so like, as you start to remove choices, you’re helping kind of narrow down someone’s decision making process, which is why we think giving people those kinds of opportunities increases conversion by such, I mean, 3X is huge, right. And so that’s something else that we’ve noticed, we’ve been at it for, we’ve been testing it for two years, but you really still learning and still kind of fine tuning and there’s some of the things we’re seeing right now. So I don’t think that you know, back to your original, I don’t think there’s anything saying that recommendations don’t work. I mean, recommendations do work 100%. But it’s time to evolve those, it’s time to, we have the technology to make it better. And everybody, and this is the part that gets me like this is my pet peeve, by the way, is that every site I see for the most part follows the same structure, the same paradigm, it’s like, let’s give them a list of infinite scroll of products to look at. Let’s give them a filtering mechanism, which is like a bunch of checkboxes of terms that I actually may not even understand like, it’s a merchandiser texting, or I’m being thrown some kind of recommender that has no idea who I am and why I’m there. It’s just a very frustrating experience of what am I going to do as a consumer, I’m going to go to Amazon because it’s easy. And I think that’s where the opportunity is for E-COM is for folks to kind of step out of that and say, like, we want to create a better experience. And how are we going to do that? We’re going to engage our customer, almost like in store like you don’t walk into the store, and they get forced down aisle one, aisle two, aisle three, aisles whatever, you can roam around the way you want. You can go talk to a sales associate, ask them questions. They’re gonna say, do you like this or don’t like that? Like, that’s what Scott was trying to do. We’re trying to recreate a more human experience.
Izach Porter: Makes total sense. It really does make sense. You’re just putting people in control more of their shopping experience? Who’s got to convert that into the Metaverse, and we can all go well, you’re gonna real.
Michael Prichard: It’s funny, you say that we are working on some things in that world, but we won’t talk about that. We’ll talk about the next Deal Closers.
Izach Porter: Okay. Maybe on the next the next show man.
Michael Prichard: The next show
Izach Porter: The next is coming for sure. So we touched on kind of AOV, and LTV. I wanted to talk a little bit about bundling. How does handling kind of affect sales and play into what you’re?
Michael Prichard: That’s actually huge actually, and we’re starting to touch that, right. So we’re starting to kind of release a few more solutions. We got caught up, to be totally honest with you. Our first solutions off our platform were a little bit hard for shoppers to understand what the heck was going on, right. And we’ve learned from that. And now we’re like doing things like we’re just, we call it Skafosify I came to the word Skafosify like, we’re Skafosifying. Oh my God must stop saying that. We’re trying to make things more Skafos. So like, we’re traditional recommender to it. And, so that we did the PLP, we added some scuffles to it. Now, we’re working on future products, bundled products and adding scuffles to it. But at the core, your core question is like, do bundles work? The answer is like, yes, they do work, right. So, but I think a lot of that could be done without AI, you can borrow but I think most, a lot of merchants know what their products are. They can start there, right. And that’s what we do now. So we are offering people the ability to like sort of create a bundled product kind of recommender on their homepage or wherever they want it. But then what we’re doing is giving agency again, giving that button to get people interactive. So yeah, I mean 100% works. It works and it’s pleasant. I hate pop up, so get me started pop ups. I’m thinking about things that work and pop ups do work but I hate them as a consumer just FYI everybody.
Izach Porter: Yeah, what do you hate about pop ups?
Michael Prichard: What I hate about him is like, I get to sites, it’s again everyone’s doing the same thing like what makes like why would izachporter.com e-commerce site be any different than michaelprichard.com site if we do the same thing, right. And so like you go to a site and within I’m gonna say 1 to 10 seconds, I get a pop up that’s either like, Hey, let me give you a discount, which is like, well, hold on now all of a sudden you’re discounting. Like, what was wrong with you guys? You know, I looked at the site yet like, chill out. We’re given a spinning wheel. Again, I get why you’re doing this.
Izach Porter: Yeah, spinning wheel.
Michael Prichard: Yeah
Izach Porter: I do hate the spinning wheels.
Michael Prichard: I hate it. But guess what?
Izach Porter: It’s always 15%.
Michael Prichard: So I will, yeah, exactly. I would but I was on like, we’re on all the Slack channels like that Slack channels. And I remember saying, like, why the hell people using pop ups? You know, they’re horrible. And then like, the people selling pop ups are like, you know what? We know, they’re horrible. But we don’t care because it works. So it does work. Everyone hates him, but it works. But then it’s like, I started thinking about like, well, I don’t know, is it? Can we do something better? Maybe offer a better experience? Let me look at least look at the site before we start popping me up with let’s like, you don’t need to give me a sale right away. Because now I’m expecting it. I’ll be honest you, I go to sites and I wait. I’m like, Oh, let’s see if I get 10% off.
Izach Porter: Yeah. Well, here’s kind of to bring that full circle back into kind of the M&A aspect that I focus on, what I’ve seen in e-commerce in tech business sales is that, the brands that can actually identify as a brand create a customer experience that is differentiated somehow, get rewarded enormously, that at the time when they sell their business. In other words, if you can execute on the concepts you’re talking about, it creates real value, not only does it improve the financial results of your company, but it creates more value for your company, the multiple on that business will be higher than a similar business that doesn’t do it as well.
Michael Prichard: Right. I mean, it’s, well, one, the cost of acquisition to get a customer is one getting way higher. So your return on ad spend is going lower, right. Therefore, you need to figure out, how do I capture that customer, not just to sell them one time back to your point, but to keep them a lifelong customer as long as possible. And to do that you need to create a brand story and experience. Essentially, you need loyal. I mean, I’m loyal to certain brands 100%, and I’m loyal to them. Because when I do you gotta have a good product, believe me. You can have like crap products, but like you have to have a good product, but then you have to have a good, it’s like, I don’t know. I mean, probably the first ones that really do that are like Patagonia. Like I mean, like, Dear Lord, like people like live and die in Patagonia, because they’ve created this amazing experience.
Izach Porter: Yeah, yeah, that’s for sure. Yeah, and if you can even just replicate a little slice of that Patagonia brand loyalty.
Michael Prichard: Oh, it’s over.
Izach Porter: You can create huge, huge value.
Michael Prichard: Yeah.
Izach Porter: So what are some success stories you can share?
Michael Prichard: I mean, so I started out with like, the Australian jewelry company. We dropped product finder initially into their front end, and they’re the ones that hit like, I think it was 105%, isn’t four days, by the way, it was crazy. 105% CVR growth, plus 85% RPV which was insane. We have done another is a great another great one. I mean, we originally started with just a traditional recommender, and we looked at the data when was it like, earlier this year, we went down there, they’re in Lynchburg, Virginia, by the way, they’re a great company. I love them. Like, they have amazing products, they’re actually good people. So to pitch them right now and you should go check them out. But we like looked up the stats. And we went in the meeting, we even look at the stats, like we’re just like, just pull up, let’s see what happens. And it was like, I forgot, I was like insane. It was like 7% conversion rate before interacting. So like, it’s pretty wild that it work. You know, it works for the most part. Sometimes it doesn’t like I said, like we’ve also seen with Michael Hill, who we’re continuing to work with right now. We ran different tests, and we did see higher conversions and things like that, but AOV was dropping. So now we’re trying to figure that out. But it’s, we still know that people are that it does work and we’re just fine tuning it. I guess the best way I’m not gonna lie to you. I mean, oh yeah, it’s gonna be great, the best thing ever, but it’s like, everybody says the same crap. It’s like the other day it’s like, look, we have to work with you and we need some time to tune it, right. Because you have different product different customers and we know it works if it’s done right.
Izach Porter: Yeah, that makes sense. So Michael, how can our listeners connect with you and skafos?
Michael Prichard: Yeah, absolutely. So skafos.ai is our domain. You can hit me up. You can go to [email protected]. You can go on our Twitter. You know, we’re on Twitter, Instagram, and TikTok apparently under been on it. And Facebook, just search for skafos.ai and talk to us.
Izach Porter: That was Michael Prichard. You can find him at skafos.ai, that’s skafos.ai. Thanks everyone for listening to this episode of the Deal Closers podcast, brought to you by websiteclosers.com. If you liked the show, be sure to rate us, write a review, press the follow button and share it with your network. And of course if you’re looking for help selling your e-commerce business, be sure to visit websiteclosers.com. This episode was edited and produced by . I’m Izach Porter. Connect with me on LinkedIn, and we’ll see you next time on the Deal Closers podcast.