Building customer profiles through facial recognition, using cameras to predict out-of-stock items, and measure visitor flow – these are just some of the ways AI may shape the future of retail.
David Bonthrone is the Chief Revenue Officer of Remark Holdings, the builder of one of the world’s largest data-driven suites of AI solutions. He joins us in this episode of the podcast to discuss use cases for AI in retail, what it means for consumers, and how retailers should approach getting started.
Read the Transcript
John Rougeux: Hi everyone, and welcome to another episode of People in Places. I’m John Rougeux, VP of Marketing here at Skyfii and your host for today’s episode.
John Rougeux: Our guest for today is David Bonthrone, the Chief Revenue Officer of Remark Holdings, a business that’s built one of the world’s largest data driven suites of AI solutions called KanKan.
John Rougeux: And David’s here to talk to us about the future of AI and retail. David, it’s a real pleasure having you on the show today.
David Bonthrone: My pleasure. I’m looking forward to answering some of your questions and having an engaging discussion about an exciting field that is moving quite quickly in front of us.
John Rougeux: Good deal. Well, in a minute I know you’re going to share some of your insights on how artificial intelligence is changing the future of retail.
John Rougeux: But before we dive into that, David, I know you have a pretty impressive background having worked with some pretty hot profile agencies and brands over the years. So, I’d love to hear a little bit more about that and the work you’re doing at Remark.
David Bonthrone: Well thank you very much. Yes, I’m based in New York and my background is really classic advertising. I started life in London as a banker and very quickly I decided I didn’t really want to do that the rest of my life.
David Bonthrone: And I went off to Australia and joined a big agency and what tends to happen is the big agency got fired by the big client. And I set up at that stage my own direct marketing business. Direct marketing wasn’t particularly fashionable but it was mighty effective.
David Bonthrone: I started to win a lot of work and consequently the business was acquired by a larger group named Saatchi & Saatchi and I went to work for them as general manager of that group. And at the time the internet came along and we started doing some basic banner work and built Australia’s first transactional website in the late 90s.
David Bonthrone: Thereafter, I moved to New York. I joined McCann Erickson. I worked on the Mastercard business and I had subsequent spells at Ogilvy in the digital brands group and I went client side for a short time as the Head of the Global Marketing for a software company.
David Bonthrone: Then I went back to the agency world and I ran a brand that was consumed 30 billion times on an annual basis but I enjoyed very much the global aspect of that. And I had another short spell at Saatchi & Saatchi where I ran the New York office for their retail practice. I did the same at McCann. And then I went entrepreneurial and the last three ventures I’ve been involved with have been sort of a startup nature.
David Bonthrone: Currently, I work at a very interesting organization called Remark Holdings. We’re NASDAQ listed. Our stock ticker is MARK, and we’ve been around for a number of years and we’ve acquired a number of digital properties that we’ve built and developed most notably VEGAS.com.
David Bonthrone: We also have built in China a pretty sophisticated AI solution, which is backed up by a very robust set of data tools behind it. Then we have some unique relationships with Alibaba and Pension where we have access to social data.
David Bonthrone: So that’s not particularly unique, any of those things. But being able to marry the two is fairly unique and we think we have a competitive advantage in that space. So, that’s where we are.
David Bonthrone: Remark Holdings, as I say, we own five or six businesses. I’m here to talk to you about KanKan artificial intelligence business that’s growing leaps and bounds in Asia-Pacific and we have interests in other areas around the world.
David Bonthrone: So we have interests in Europe and also in the United States from different types of retailers who are facing different challenges be they a convenient store or a mass retailer.
David Bonthrone: So, I’ll be delighted to talk a little bit about that. Our organization from people wise we’re headquartered in Las Vegas where we have 200 folks and we have an office in Los Angeles. We have a significant office in Chengdu, the world’s largest, the 16th largest city in China, which is our development office for our AI business.
David Bonthrone: We have a business, an office in Shanghai and also in Shenzhen. I live and work in New York and we have one person here. I’m often on a plane as one can imagine and talking with customers here in the agency board and in Europe. That’s it.
David Bonthrone: And then I think what we’ll do now is probably get into the AI component of our KanKan offering, right?
John Rougeux: Yeah, that’s right. Artificial intelligence is one of those topics that I think people are pretty fascinated by. But I think one of the challenges is that members who understand the technology really well it’s a little abstract.
John Rougeux: So I would love to hear your definition on what artificial intelligence is and what it isn’t especially in the context of retail.
David Bonthrone: That’s a great question and there is an awful lot that’s been written about artificial intelligence. And I think we’re still looking for the very clear 10-word definition of the said discipline that would encapsulate everything.
David Bonthrone: From our point of view at Remark and our AI capabilities at retail, we are essentially on behalf of retailers looking at consumer behavior via facial recognition, object recognition, gesture recognition, which involves a camera, okay, on your way to the store, inside the store and as consumers check out. Okay?
David Bonthrone: So, we’re looking at behavior and we are marrying that behavior with data. Now that’s where people get a little confused. So, let me give you a real life example of how marrying a camera with data and facial recognition might work.
David Bonthrone: Imagine if you will, that you’re walking to a Big-box retailer in China or wherever in the world and we do this with a progressive retailer in China named Lotus, and there’s a press release that’s out about this on our website, where somebody is walking towards the store and as they approach the store there is a fairly significant screen there that recognizes that particular individual because of the facial recognition software. Okay?
David Bonthrone: After it recognizes that individual, it knows based on previous purchased history what he or she has purchased at the store. So at that point it sends or presents an immediate personalized message to that particular individual about items that are in the store perhaps on special, perhaps available, that have been newly introduced to the store or newly launched so that the consumer can look at that information and act accordingly.
David Bonthrone: What we could also do, okay, is at a later time after they’ve left the store environment is give them an offer to return back to that retail environment. So, therefore, closing the loop on the engagement from a consumer point of view.
David Bonthrone: So that’s one … When I talk to senior executives at retail, they say to me … I’ll give you an example, “I’ve heard about this AI stuff.” So that’s one of the examples that we can use. There are others. So it depends what type of retailer you are and it depends what your … what insight you’re looking to uncover or perhaps what challenge you’re looking to potentially overcome.
David Bonthrone: So, for example with the big convenient store chain that we work with in Thailand, they are very interested in using artificial intelligence, again it’s a camera, okay, to predict, tell them when there are items that are out of stock. Okay? So, instead of having somebody walk around the store and bend over and look down there and stare at the other stock, we can do that through camera technology and we can report back in real time to management at a national, regional or more likely at a local level to identify and say, “Said particular stores, 3, 4, 5 and 6 in that particular area are suffering significant out of stock levels. Why is that?” That is because the store manager isn’t paying enough attention to that. And we can fix that problem pretty much immediately.
John Rougeux: Right.
David Bonthrone: There are multiple, multiple applications for this artificial intelligence. It’s really like marketing. There are multiple things that marketing can do. New cast buyer acquisition, increasing the basket value of customers and such like.
David Bonthrone: I think the point is from management at retial is to identify what the business and the clear objectives behind AI might be. One of the other things that we want to cross is we use the facial recognition and gesture recognition to build very detailed profiles of consumers integrating all of the data that can be used for future prospecting and messaging and stuff like that.
David Bonthrone: I think it’s fair to say that at retail there were many legacy systems and let’s say disconnected sources of data that do not add up to provide a holistic view of the consumer, okay? That’s on the consumer side.
David Bonthrone: On the operations side, this recognition technology as well can help store managers ensure that shoppers are visiting particular areas of the store. That’s something we do with another retailer. We’re watching traffic flow inside the store and we are seeing that a very low percentage of total shoppers are actually visiting a particular category or aisle.
David Bonthrone: So, what does the store manager need to do in order to fix that? Is it a price issue? Is it the fact that the aisle is not a good experience for the consumer because it’s not appealing dirty, there’s maybe not enough going on there. We can use that technology to identify what the barrier is so that the retailer can fix that.
John Rougeux: All right. So you mentioned a few different examples there. You mentioned a marketing example of identifying shoppers and personalizing some offers. You mentioned an operational problem being solved with the store shelves and looking for potential outages.
John Rougeux: And then I think you mentioned more of an insights, so kind of the behavioral analysis type situation where you’re looking at customer flow and in-store analytics on movement and sections of the store or venue that are being visited or not visited.
John Rougeux: I’m really curious about this first example, David, because it goes back to something you talked about earlier in your marketing career. You talked about how direct marketing wasn’t something especially popular at the time and I am curious what the reception on the consumer side has been to having their faces recognized as they come into the store. I know you’re doing that in China right now. So, what have shoppers thought and said about that experience?
David Bonthrone: That’s an excellent question. Just on the direct marketing side it’s like direct marketing has always been frankly hugely popular to the smart markets or it’s never been particular sexy though.
David Bonthrone: So, it wasn’t something that was adopted by the big brand advertising agencies but let’s look at the whole advertising world. Now the growth has come from everything that CRM and the 30-second TV series agencies are having their challenges.
David Bonthrone: In answer to your question about the data, how the consumers feel about that, in many situations it’s rather like the internet, it’s very wise to provide an opt-in scenario. So we do that and we work with the retailer to do that. If the retailer does not want to do that, then that’s a different discussion and we’ll pursue that accordingly.
David Bonthrone: Rather like consumers are prepared to, but rather like consumers in many instances are prepared to give up some data for loyalty programs or membership rewards programs and stuff, we provide the facial recognition just as another identifier. So if you can look at, see somebody through a screen, through a facial recognition technology and that then links to their past purchase history, all of their loyalty program that the retailer may have, is just another identifier. So it’s a link between what is an existing situation already because the retailer has loyalty data on certain consumers and we just need to dig a little bit deeper here in the world of Tesco in the UK, for example, and their relationship with Dunnhumby with whom they work very closely to explore that. That’s and pretty much most retailers around the world have been developing loyalty plans or programs of some yoke.
David Bonthrone: What we’re doing is providing facial recognition over the top of that as another identifier of that consumer enabling us to pull that data and have access to it and then in real time send a message to said consumer on digital device about a particular offer that might be in the store.
David Bonthrone: So, if the retailer wants us to have an opt-in function at facial recognition level, we can absolutely do that. However, in most instances, the consumer has already provided a lot of that information because they’re likely to be part of the loyalty program.
John Rougeux: So are you finding that there is a level of education that has to take place with consumers about what the technology is, how it’s used. If there’s an opt-out scenario how that even works or is this something that they get pretty intuitively?
David Bonthrone: Yes, there is a level of education and we would never, ever want to be seen to breaching any data standard that we’ve been adhered to. However, we all know what consumer behavior can be like when they see the next shiny best thing, okay?
David Bonthrone: So, some of the stuff that we are doing at retail in China, I’ll give you another example, when consumers leave the store in China after they’ve checked out, there’s … If they have some dwell time, let’s say they are waiting for somebody else to come out of the, your kids to come out of the store or somebody is having a coffee or whatever. There are screens on the outside of the store and sort of sitting area that are interactive, driven by facial recognition and they can spit out a coupon for you to return to the store where you can, digitally obviously, for future engagement.
David Bonthrone: So we are using artificial intelligence and facial recognition, gesture recognition, observation, all of that stuff, to further enhance the CRM model or in deed close the loop with the consumer at retail. And I think that’s a very powerful thing using technology to do that.
David Bonthrone: One of the things that we look at in retail is the free standing insert that’s been around for many years. It works at a certain level and that’s why retailers continue to use it year in, year out.
David Bonthrone: But, using facial recognition to send a message to a consumer as they are entering the store in their digital device is far more effective from the response point of view, far more responsible from a sustainability point of view, preserving all of those trees, and it’s, I’m not saying it’s the end of the FSI, but a lot of us who’ve grown up in retail and shopper marketing have challenged the effectiveness of the free standing insert that drops out of the newspaper every day.
John Rougeux: Right, right. Yes. So in many ways there’s a potential to provide a better experience because it’s more real time, it’s more personalized, it’s more accessible, you aren’t necessarily creating that negative experience of getting what may be perceived as a piece of junk mail in the newspaper that goes straight to the trash can.
David Bonthrone: Yeah, absolutely. And guess what? There is a market for that. There are people who’ve had that all their life and they are not going to change. We respect that.
David Bonthrone: But in fast-developing, technology-enabled, high digital consumption markets like China, by the way, because they’re ahead of the US when it comes to artificial intelligence, we see behaviors that are different.
David Bonthrone: I would invite anyone to, if they want to explore further some of the work that we’re doing in China, what we do with our partners down there, is probably a year or two away of being done here in the United States because of the various things you’ve spoken about. And actually it’s very exciting.
David Bonthrone: One of the concepts we are working with a European retailer at looking at who’s developing potentially a staff-less store and having artificial intelligence solutions drive that.
David Bonthrone: So, we use facial recognition to validate you when you go into the store, okay? We will use facial recognition to prevent underage kids from buying restricted categories like alcohol and tobacco. Okay? And then after you’ve scanned your items, okay, at checkout, we will verify your presence of we’ll verify your identity through facial recognition allowing you to leave the store.
David Bonthrone: Then the future might look like we have small little robots doing shop replenishment at low customer time, i.e. when there’s few customers in the store or when the store is closed over night. And we’ve seen that. A couple of retailers are piloting that and trying that.
David Bonthrone: So these are all very exciting things for the future. Of course, underpinning a lot of this is the fact that retailers today are traditional brick and mortar retailers are challenged by the whole online, eCommerce explosion.
David Bonthrone: So, they’re seeking always to increase that top line by making more sales while at the same time seeking to remove some costs from the business. Now, particularly at Big-box retial, these stores are very labor intensive.
David Bonthrone: So, if some of the stuff can be done by a machine, the savings there are potentially significant. We’ve identified without naming any names that some of the initiatives that we’ve undertaken at our retail level have shown a potential cost saving of between 20% to 30% at a store level basis.
John Rougeux: That’s a pretty sizeable number to consider.
David Bonthrone: Which is pretty significant, and the way … One of the ways I look at this AI technology is let’s go back 10 years ago where the whole outsourcing/offshoring phenomenon came, and these enormous companies in low cost markets were grown. Groups like Withrow and Infosys in India and others in different areas, and they took the non-core functions of certain organizations and outsourced them. Be that call center or whatever it happens to be.
David Bonthrone: And what we’re seeing now is a slowing of that but a replacement of that via technology, not just that lower labor cost.
John Rougeux: So there’s a … I think that brings up a good point for someone who’s potentially looking at artificial intelligence. On an earlier episode of the show one of our guests was talking about how when you’re looking at new technology you really shouldn’t pursue it for its own sake and fall victim to shiny object syndrome.
John Rougeux: So, in the case of artificial intelligence if a business is looking at that category of technology and thinking about maybe exploring that further or doing a trial, doing some tests, things of that nature, how would you suggest that they approach that discussion to make sure that they’re actually going to get value from that exercise?
David Bonthrone: Well like I said earlier in the conversation, I think you have to set certain objectives, okay, which identify what the bottlenecks or what the challenges in your existing business are and set certain objectives around that and work with your AI partner on a quantitative basis so that they can improve the financial impact of said initiative.
David Bonthrone: Now, with a lot of these technologies we’re all aware there’s some upfront set up costs but the value is realized over a period of time with scale. So that’s why one needs to set quantitative objectives up front on a four or five store test perhaps and then you can project and roll that out over 200, 500, 3000 stores, however many it happens to be.
John Rougeux: So start with some quantitative objectives, do a test, see what kind of results you get and then go from there.
David Bonthrone: Absolutely, you are absolutely right, absolutely right. So, we are working with an incredibly progressive retailer in China and they’re a mass retailer, Big-box retailer, they sell everything from grocery all the way through to odd ware and personal care, rather like a Target or here in the United States, incredibly progressive.
David Bonthrone: And we’ve worked with one store with them at the moment. Okay? That’s our proof of concept, that’s our pilot, and we’ve looked at some very clear metrics and we have some very impressive early stage metrics on the consumer side and on the B2B side in terms of operations.
David Bonthrone: But we can extrapolate and look at okay, fine, so if we got X more consumers over Y period spending a certain amount of more money as a result of our AI initiatives, when you look at that behavior over the course of a month, we look at that behavior over the course of a year, or a quarter, we factor in some seasonality components and then we look at that over 1000 stores or 500 stores or however many that the retailer has and then we look at the cost of the upfront installation and the ongoing SAS license fee, it becomes financially very feasible and it’s like, “Yes, this is a very obvious way to improve performance at a local, regional, national store level and at the same time to eliminate some unnecessary costs.”
John Rougeux: Good stuff. Well, David, you have shared a ton of valuable information about artificial intelligence, facial recognition and retail today. I really appreciate all of your expertise. Do you have a favorite book, website, author that you recommend?
David Bonthrone: Too many to mention. Too many to mention. I read a lot of newsletters, AI VentureBeat is one that I like tremendously. There’s another couple that I read. But I read a lot.
David Bonthrone: But there’s MediaPost and MediaNews. And if you look at MediaPost, there’s a gentleman there called Chuck and his last name escapes me. But he writes some very interesting stuff about artificial intelligence.
John Rougeux: Good stuff.
David Bonthrone: And all things tech and retail.
John Rougeux: All things tech and retail, okay. All right. Well, David, if one of our listeners wants to get in touch with you, what’s the best way for them to do that?
David Bonthrone: They can email me. My email is Dbonthrone, that’s B-O-N-T-H-R-O-N-E @remarkholdings.com or yeah. They can find me on LinkedIn or they can go onto the Remark Website and yes, get a message to me that way.
John Rougeux: All right. Well David, thanks so much for being with us today. It was great having you.
David Bonthrone: Thank you very much for your time and if anybody wants any further information I’d be delighted to provide them with that.
John Rougeux: Good stuff. All right. Well that’s it for today’s episode of People in Places. If you like what you’d heard you’d be doing our team at Skyfii a huge favor by leaving us a review on Apple Podcasts. To hear more just search for People in Places in your favorite podcast player.
John Rougeux: To get notified of new episodes, follow Skyfii on LinkedIn or subscribe to our newsletter at Skyfii.com. That’s S-K-Y-F-I-I.com. Thanks for listening. I’m John Rougeux with Skyfii. We’ll see you next time.
People in Places is a podcast dedicated to helping today’s shopping centers, retail outlets, airports, museums, universities, and other physical locations optimize the experience of their visitors. Get in-touch: email@example.com. See all current episodes on our website here.