Read the full transcript
Great, thanks for the introduction. And again as mentioned, this presentation might be a little different. In that I’ll be looking more three years out, five years out, even 10 years out, to see how AI and IoT and other emerging technologies are going to reshape the format of brick and mortar and online retail and their future blend. So I’m here from CB Insights, a technology company based in New York that uses machine learning to track high momentum start ups as well as broader trends in emerging technologies, patent filings, strategy innovations and more. We work with some of the, we work with major corporations including some of the biggest retailers and food companies around the world. And we believe in facts over feelings. We also have a series of innovation focused events geared specifically towards C suite executives at one billion dollar plus companies. If that’s of interest to any of you in the room, feel free to track me down afterwards. And again I’m the lead CBT and retail analyst. I have a weekly retail tech newsletter. Please, it would be great if you signed up. So let’s get started. I think all of us in this room know that brick and mortar retail has been in a bit of a tough spot over the past few years. We’ve seen bankruptcies multiply, particularly among these sort of mid-priced mall based apparel and accessories retailers. And beyond bankruptcies, we’re seeing thousands of store closings around the country as companies like Addidas and Kenneth Cole refocus their investment on e-commerce, as retailers realize they’re over built in an age of e-commerce. And we’re likely to see this contraction continue, especially since the U.S. has more retail floor space per capita than any country in the world by far. So to kind of balance this out, we’ll likely see thousands more of store closures over the next few years. But this is not to say that brick and mortar retail is dead, rather as Lisa Landsman, the former President of debt.com put it very wisely, it’s pure play retail that’s dead. Companies need to be both online and offline, as well as in some new channels that we’ll discuss later on in the presentation. And I think we know, I think this is intuitive to all of us here. We know that while traditional retailers are closing stores, start ups that launched as online only are opening brick and mortar stores. Most notably, Bonobos and Warby Parker, but also all you have to do is walk around Soho a little bit and it’s like a playground for flagship stores of online only startups like Casper, Allbirds, and more. So we’ll likely see this, this kind of happy medium as online only retailers open stores and brick and mortar retailers start to close their doors. And AI in the future of all of this, this happy medium between brick and mortar and e-commerce, this future of omni channel blending, is going to be omni present. We know that AI is hot. It’s why we’re all here today. According to CB Insight’s data, AI startups have raised over 40 billion dollars since 2012 around the world, including nearly 16 billion dollars last year alone. We also at CB Insights put together this list of the 100 highest momentum AI startups in the world. If you’re interested in this again, find me afterwards. I know the screens are a bit small. But throughout all of this, the most important thing to remember is that AI is not a silver bullet. AI is a tool for analysis. And these algorithms require huge data sets to train on, to optimize over time, and then to run on in order to provide any value. And when we think about who’s going to win the retail wars of the future, especially in the age of AI, it’s going to be the companies that have the most data. Amazon of course is a great example. They must have one of the, maybe the biggest retail data set in the world and on top of that, they can build some of the most sophisticated AI tools and algorithms. So as retailers prepare for this tech enabled, AI powered future, data collection is going to become more and more important. And brick and mortar stores can become valuable foot soldiers in these retail wars, these data collection wars of the future. But as we know, brick and mortar retail footprints are shrinking, so going forward we can start to think of new ways to extract the most data out of smaller and smaller brick and mortar real estate footprints. I think again we all know, it’s getting harder and harder to measure sales per square foot in a retail store in any sort of real meaningful way. Since so many customer journeys start online and end up in store or maybe people browse in store and then buy online later. So I’d like to encourage you all going forward to start thinking of this model of data per square foot, rather than sales per square foot, as a way to sort of conceptualize this strategy of extracting the most customer data out of the brick and mortar real estate that’s still available. So how can we use brick and mortar retail stores to really engage with shoppers meaningfully, give them great excuses to give up some of their own personal data that retailers can later use. Technology startups are huge players here. So let’s take a look at some of the more interesting examples of new technologies that are increasing the data per square foot metric. The major theme is that all these new technologies are focused on gathering data directly from consumers, rather than relying on the self reported behavior that retailers have traditionally had to rely on in the form of focus groups. So now that we have IoT, AI, machine vision, that we’ll go over, we can see what shoppers are actually doing, rather than what they say they’ll do. We don’t have to ask, you know, shoppers whether they buy, whether they eat healthy. We can actually see them going into that convenience store for that 10 a.m. frozen burrito. So some of these startups observe from the background. Collect data without consumers necessarily even knowing. Trax retail, which has raised nearly 150 million dollars, based in Singapore, is one example in the CPG space, which helps brands like Coca Cola monitor the performance of their products and shelves in third party retailers. Trax also recently acquired the store observation unit from Nielsen, which they’ll integrate into their machine vision tools to help retailers, I’m sorry, CPG brands access even more and more data. We also see AI being used to capture online and offline purchasing data, and then use that data to adapt prices in real time, both in store and online. So for example, in a pretty interesting just recent partnership, Altierre, which is a IoT startup partnered with Eversight, an AI startup, to create digital tags for products in stores that change the prices listed for products in stores based on real time purchasing data, real time promotions that retailers might want to roll out in real time. And other startups offer a whole range of data collection methodologies to retailers. Farfetch, for example, is a luxury e-commerce startup, but they recently piloted a modular store of the future platform, which they plan to sell to retailers. So brick and mortar retailers can kind of pick and choose different tech enabled data collection strategies from Farfetch to just sort of plug and play into their existing stores, including universal logins to track visitors, smart mirrors, online integration, online offline wish list integration, and more. Other data gathering methods are more up front. So for example, we’re beginning to see in store robots greet shoppers in certain areas. I don’t know if any of you here have seen one of these robots in person. They’re pretty cute. They’re about four feet tall. This one is called the Pepper robot produced by Softbank. And they are already active in over 10,000 venues around the world, greeting shoppers, creating you know a fun, kind of gimmicky experience for shoppers. They do boost foot traffic in the short term at least in places they’re installed since shoppers want to come check this out. It might not be the best long term strategy for attracting foot traffic since the novelty will wear off, but while these robots are greeting shoppers, they’re also gathering visual data on shopper demographics, recording the questions that shoppers ask. All this data that would be lost in a world of just human greeters. And then beyond these humanoid robots, we’re also seeing retailers like Walmart, roll out robots focused on inventory management. So these pilots seem to be a success. Walmart recently expanded their pilot with Bossanova robots. They’re right here. These are in over 50 Walmart stores today. These robots aid in restocking shelves, but also are monitoring what shoppers are doing in the stores, tracking what products get picked up, put back, sold out, more quickly. And so these are huge new data streams. And speaking of these fun robot experiences, when you start collecting data directly from shoppers in this more up front kind of face to face way, where shoppers are aware that you’re collecting data, you also, you have to offer a better, you have to pay the shopper for that data in some way. Be that in money, be that in product promotions, be that in a better experience. So very directly, Vengo is one example of a startup that sets up these little digital kind of automated tablets in stores and asks shoppers some personal survey questions, and then in exchange for answering those questions, for giving up that data, shoppers get free product samples. So here Vengo helps retailers pay shoppers directly for their data with free product samples. We also see startups that help retailers pay their shoppers in better experiences in exchange for their data. So for example we see a number of startups offering plug in play, free guest wifi platforms for stores like Zenreach, which has raised 80 million dollars. Cloud4wi, which has raised 15 million dollars. So this improves the experience for shoppers, since they get free guest wifi, but also helps the retailer collect data on in store browsing activity, the shopper’s email address, can track repeat store visitors every time their phone comes back near the store in the range of the wifi, and more. And then even in kind of a more eye catching better experience, we see startups like Perch, which sets up digital interactive tablets in stores that can highlight new products and calls attention to the products that are on offer, but also is collecting data on which items the shopper picks up, which items they put down, which items they ultimately buy. So this is very valuable. Of course a traditional shelf cannot, you know, no matter how flashy and eye catching, cannot be recording data on shoppers interactions with the products. And Perch does say they work with Kate Spade, Levi’s, other major clients, and they say other tablets boost sales by 30 to 80 percent. So Perch makes the digital hardware, Modiface makes the software. Both Perch and Modiface are in Sephora. If you’re vising New York and interested in seeing a great examples of brick and mortar retail, I’d recommend checking out the flagship Sephora store in midtown, where you can see all of these kind of augmented and virtual reality features. These interactive tablets that let store visitors access makeup tutorials, see how different looks would look, different make up styles and colors would look on their face. So this is a very engaging, very exciting experience for shoppers, but is also collecting so much data on which products shoppers try on, which looks they pair together, what products people of different complexions might gravitate to. They can then help Sephora and other retailers better design their promotions and also better design the products themselves. Maybe thinking of new styling tips that people are trying in stores that they haven’t thought of or things like that. Further out, we can see shoppers actually getting truly personalized products based on their data. For example L’Oreal has looked into a method, they filed a patent and also created sort of a pilot product that would observe shoppers through machine vision, and then print out on the spot a customized makeup blend based on their skin tone. So this of course helps them collect data and also would create a fun experience for shoppers and a real personalized product. In China we already see facial recognition technology much more developed and much more common with major players like Faceplusplus, which has raised over 600 million in funding. And Sense Time, which has raised over a billion dollars and is backed by Ali Baba and other major players. So these startups use AI to create facial recognition platforms that aid retailers. We do see them even aiding police departments in China, which is interesting. But also businesses, other companies, and these offer benefits for shoppers. They’re fun to use. Most notably, maybe we’ve seen recently, that KFC in China has been piloting a smile to pay system where shoppers, you can kind of see it a bit here. Shoppers look into the screen, smile, their faces register, that’s connected to their personal profile, and the payment takes place then automatically through their Alipay or WePay account. So features like this, which make the payment experience easier and more streamlined for shoppers are sort of habituating people to this facial recognition technology, which then can be integrated with other retail operations. And of course collecting massive new data streams for retailers, everything from demographic data to shopper emotions while they’re in stores, how shoppers are emotionally reacting to different displays, and even I think this is perhaps a bit distasteful, but they’ve recently piloted giving people beauty scores. And then using that to promote different makeup products or even to integrate that with dating websites. So you know, a lot of interesting things going on over there. So now that we’ve talked about how new technologies can help retailers collect more and more data to power AI platforms through their brick and mortar stores, we’ll take a look at use cases for AI online in the e-commerce world and how AI is both collecting and analyzing new data streams as well as personalizing e-commerce platforms. As I mentioned before, one of the key themes to keep in mind is that AI helps retailers analyze what shoppers are actually doing. Not relying on self reported behavior. So most directly maybe, we see startups like Brandwatch, which has raised 63 million dollars, works with Unilever, IKEA, and other major companies, scrape social media conversations and then use natural language processing to analyze sentiment, what people are saying about your brand, again what they’re saying, what they’re sort of actually talking about with their friends and peers not with their self reporting to the brand. We of course see startups like Dynamic Yield, which collect shopper data and personalize websites in real time. And then also AI to track shopper behavior across devices and then help retailers personalize their promotions and websites in that way. For example, Appier, which is based in Taiwan, works with Carrefour and some other major brands around the world. AI also powers personal, sort of personal shopper profiles, which make the online shopping experience better, easier for shoppers, and also gives retailers access to huge new streams of data. So True Fit is one of the most notable examples here. They work with Micheal Kors, Blommingdales, Ralph Lauren. They’ve raised over 100 million dollars. And the way True Fit works is it creates for the shopper a personal fit profile that travels with you across websites, across apparel websites. I know the screens are a bit small, but on the right you can see when you log in to a website, that when you visit an apparel website that uses True Fit, it just asks you for your height and your weight very easy. And then recommends what the best size is to buy from this particular retailer in this particular product. So this makes the sort of experience of buying a piece of clothing online much more trustworthy for the shopper, and also True Fit’s major value add for the retailers is demographic and body shape information on millions and millions of shoppers around the world, as well as their preferences for styles and certain looks. After the point of purchase, AI can also help to personalize loyalty programs and outreach to keep that shopper engaged. Signpost, also based in New York, is one example here that works with brick and mortar retailers to personalize a cadence of emails encouraging the shopper to leave a review of your business online. And they say the businesses they work with see significant lifts in their Yelp and Google review scores. Starting to look even more forward into the future, we see how these huge data sets that are being collected by AI online and offline will help retailers design new types of retail formats all together. For example, digital payment technology, which is really an e-commerce innovation, are supporting new brick and mortar formats in the form of fully automated retail stores. The most notable example that many of us in this room may have heard of is Amazon Go. The fully automated cashier free convenient store that Amazon recently opened in Seattle. So this store uses AI and machine vision to track shoppers throughout the store and then charges them automatically to their Amazon account as they walk out the store. So you just pick up the product, put it in your pocket, and walk right out and you’re charged automatically. It did take Amazon a bit longer than they expected to get this off the ground, but the Amazon Go store in Seattle is now open and operating. And just like facial recognition, this system is also much more mainstream already in China because the mobile payment ecosystem is much much stronger in China than it is in America. And China are already very used to using mobile payment apps like WePay and AliPay. And these are being used to support these fully automated human free convenience stores like we can see in the pictures here. The way these stores work is you scan your mobile payment app in order to enter the store. So from the very moment you enter the store, the store knows exactly who you are and has access to your payment information. You then pick up your items, scan you phone again to pay and leave. These are very, so these are small scale sort of convenience stores that are popping up in cities around China that are able to take the human cashiers out of the equation by relying on this kind of initially e-commerce focused digital payment technology. And as all these brands collect more and more data, by linking every brick and mortar purchase to a digital payment account, they can start to be predictive about what people will buy, what objects, what products, where and when. So one model that is beginning, is emerging, and will likely strengthen in the future is the dark store model. This brings together the themes of shrinking and retail real estate. The combination of online and offline data, and also making the purchase experience as easy as possible. The dark store model refers to basically delivery only stores. So these are small scale, small scale pop ups, small scale warehouses that don’t allow any shoppers and they only store products and retailers use them to support one hour deliver in cities like New York and San Francisco. So they don’t have to pay, you know, a lot of retailers today use fulfillment from store to support faster delivery for e-commerce purchases. But with a dark store model, you’d no longer need to pay for the real estate of a full consumer facing store. You only need a small space that stores products that can then support delivery only, I’m sorry, fulfillment of online orders. Two startups here, Common Sense Robotics and Darkstore. As we start to see autonomous delivery drones take off, these of course are still in their relatively early stages but are hitting the streets in Arizona, San Francisco, Las Vegas. 7-Eleven even delivered recently a slushy by drone in Las Vegas. Not sure if anyone saw that, but we can see these autonomous delivery drones start to integrate with these mini urban warehouses. So the entire process from when someone buys something online to the time it’s delivered to their door could be completely automated. That order would be sort of picked and packed by robots in the small delivery only store, handed off to a delivery drone, and brought right to the person’s door. And as this purchasing experience gets more and more convenient, retailers will also be collecting more data about people’s day to day purchasing habits, as well as their general lifestyles and where they live. Looking far far out, thanks to Amazon for sort of stretching our minds and making us more creative, but we can see that Amazon has even looked into the idea of flying warehouses that would use data on purchasing habits to move the warehouse itself toward areas of high demand in real time. This comes from a patent filing. In the patent filing, for example, Amazon sites the example of a football game. They might know a football game is happening, they could fly their warehouse blimp toward the stadium, stock it with jerseys, hot dogs, things like that, and then shuttle those products down by drone into the stadium. So we see, you know, I’m certainly not saying this is something that Amazon is going to build. They did patent it, they might not build it. But it shows how bringing together these new streams of customer data in real time can help retail become more adaptive. And again this chart comes from Jeff Bezos, but just we see how the virtuous cycle continues in terms of retailers collecting data, using that data to make the shopping experience more and more convenient for people, and then as it gets more convenient, people will shop more, retailers will collect more data, the cycle continues. Looking perhaps even further out, we see that smart home technology can kind of complete this circle, complete this blend of online and offline purchasing while increasing data collection. So we see this starting a little bit on the form of automated kiosks that are installed in the lobbies of apartment complexes or office buildings. There’s been a few startups. For example, Bodega.ai on the right, which did get some bad press, but their goal was to set up AI powered mobile payment accepting automated kiosks that would use track purchasing data, use that to stock the small kiosks with things they knew the people in this apartment complex were most likely to buy on a given day at a given time. And then to make that shopping experience much more easy, we also see the picture on the left comes from WeLive, the co living space operated by WeWork down in Fidi, and we see that they do stock products in the building itself in the apartment building, which people who live there can buy quickly through a mobile payment system, which gives WeWork access to all their data about purchasing habits. Amazon is operating here as well. They recently spent 1.2 billion dollars to acquire Ring, a smart home startup. Ring is a connected smart security system. So it’s an interesting buy for Amazon, but what seems to be Amazon’s strategy is that they’ll use this to make their grocery deliveries or e-commerce deliveries in general more convenient for people. Since people that have the Ring smart home security system can sync that with their Amazon account and give access to Amazon delivery people. To enter their home and bring groceries even directly into their refrigerator while they’re out. So this is helpful for the shopper, but also gives Amazon a really strong foothold in people’s actual homes. Walmart has considered similar strategies as well. They have partnered with a smart lab startup called August. And they patented a solution, which would use, turn the e-commerce package itself into a sort of a key to open people’s doors, people who had smart locks and give delivery people access to their homes. Again collecting more and more data. Voice is a huge part of, the future of the smart home as well. Here we see the chief digital officer of L’Oreal says that he believes Voice may be as big as the entire internet. So this opens both new challenges and opportunities for retailers and CPG brands. Of course ordering things by voice means you’re less likely to go visit the store or even to go visit an e-commerce website. So brands do have to be preparing for this future. One in six Americans already own a smart speaker, with Amazon Alexa being the most common. And as Amazon increases the strength of its private labels, of course as you might expect, it favors its own private label products in its voice ordering systems. So for example if you order, you want to order some detergent or order a new shirt, through your Alexa smart speaker, Amazon is much more likely to ship you its own private label product rather than suggesting you take a look at a Procter and Gamble brand or something like that. So this is a bit of a threat that retailers and brands have to prepare for. And to drive home the point of just how much data these new smart home technologies can collect, we can look at a recent Amazon patent. I know the font is a bit small, but it talks about using visual screens in the home, which could be similar to the smart lock, the smart home security system that we saw earlier, to track people’s emotions, observe people’s outfits, observe the types of furniture they have in their home. These things are all listed explicitly in the patent. And then to use that to better market toward these people. Walmart again is also preparing for a smart home based future in which people are less likely to visit stores or even e-commerce websites. For example, Walmart has looked at a smart trashcan that would record when you throw away a product, an empty wrapper, and then automatically order you a new one. Procter and Gamble, a very traditional consumer company, has looked at this, a very similar system as well, which would attach connected IoT tags to products and automatically reorder the customer products when the product is used up or wears out. Then to take this towards sort of the ultimate conclusion, Walmart has been looking into automated in home store fronts. So these would be connected AI powered kiosks that might be in a person’s actual home and would stock unsold goods, kind of like a minibar in a hotel, and then charge the shopper automatically when they actually took a product and used it. So this is preparing for a future in which retailers understand their customers so well, but from both online and offline data, that they can predict what products they might like in their home and then charge them automatically without the shopper having to visit a store or a website. So we’ve been through some pretty wild, out there examples. I hope this was helpful. If I can leave you with a few thoughts, it would be that using brick and mortar stores to collect data will become more and more important, especially as brick and mortar footprints shrink and especially as more and more retailers are looking at things like experience centers, pop up shops, that focus less on immediate sales and more on long term brand building. Data collection is more and more important. Secondly to reconsider the reliance on self reported behavioral data and focus on gathering data directly from, by observing shopper’s activities directly. And then looking at perhaps 10 years out that smart homes will be the ultimate erasure of that online offline divide. And this is a challenge for which we should all be prepared. Thank you.
Zoe Leavitt, Senior Analyst at CB Insights, dives into the ways tech startups are leveraging AI, IoT, virtual reality, and more to integrate offline and online experiences and revive physical retail.