25 Jan #NRF2018 – An interview with Arun Nair (RetailNext): «Introducing the World’s First Retail Sensor with Onboard Deep Learning-Based Artificial Intelligence»
We know that Artificial Intelligence (AI) delivers unparalleled insights for eCommerce, from personalized product recommendations to product pricing optimization and precise ad targeting, to name few of them.
But what about AI inside the store?
RetailNext, the worldwide expert and market leader in smart store retail analytics, just announced the second generation of its Aurora all-in-one IoT sensor, featuring the retail industry’s widest field of view and its first edge sensor with integrated deep learning-based AI.
I had the chance to sit with Arun Nair, CTO and co-founder of RetailNext at NRF 2018, Retail’s Big Show & EXPO, presented by the National Retail Federation, January 14-16. Cold weather in New York City that morning for the team (RetailNext is headquartered in San Jose, Calif) but a booth crowded after the announcement made earlier about AI and a collaboration with Facebook. An excerpt of our conversation follows.
►Retail Next is 10 years old. What was your idea when launching the company in 2007?
Arun Nair: The reason we founded RetailNext was to say to the skeptics, especially to those in Silicon Valley, that stores are alive and well, and will remain relevant. Retail will be optimized, will shrink, and will need to be much more efficient, but the store experience is not going away. And the way to be efficient is to use data. The success of e-commerce is all data-driven We need to do the same for physical stores, to collect data to understand shoppers better. Which of our stores are going well, which of our stores are not doing so well, what can we do better, what can we improve? That’s how we came to start the company.
►How do you do that?
We use computer vision primarily. Our Aurora all-in-one sensor, the first to integrate stereo video analytics, Wi-Fi, Bluetooth BLE, a beacon and 30 days of high-resolution onboard video recording into a single device, can track anonymously and, like in a website, tell the retailer the shopper journey in the store: how many shoppers are visiting the store (with staff exclusion), information about demographics like gender, age groups, and so on, shopper behavior classification and what shoppers are doing: looking at an item, touching an item, are they trying on shoes or other products, looking for help, taking items, or just standing there confused.
Anonymous and aggregated data about shoppers are actually very useful. Using the data, retailers can do a better job in serving and selling: they can optimize the layout, assort the product mix, and educate store associates to better serve shoppers. Exactly what e-commerce retailers can do online, brick-and-mortar retailers can do in-store.
‘‘Exactly what e-commerce retailers can do online,
brick-and-mortar retailers can do in-store”.
►You have made three big announcements during the Retail’s Big Show.
The first is Mobile Engage™ a solution that offers an app-like user experience, and tenables the retailer to engage shoppers on a one-to-one basis, even without an app. A customer entering a store can log on to the retailer’s Wi-Fi and register, giving his email address, for example. If the retailer has no Wi-Fi, he can send to the customer a link by text message. Through engagements over time, the retailer begins to develop profile of the customer with her preferences.
Of course, the shopper needs a good reason to engage, like receiving a deal, a recommendation, or downloading an app for a special offer. Many customers like to get deals. Then, the next time this customer enter in the store, he will be invited to log in, and be redirected to a personalized URL, instead of going to the generic website of the company. He will be recognized and will receive personalized recommendations, offers and content based on his past history. Think about it: how many times do you go to a store and the store has no information on you, no context?
”Mobile Engage™ uses machine learning algorithms that get smarter
with each and every interaction between brand and shopper”
But there is more: Mobile Engage™ uses machine learning algorithms that get smarter with each and every interaction between brand and shopper, allowing for increasingly relevant and personalized communications at the right time. Mobile Engage™ doesn’t replace email marketing and promotions, it works with the campaigns and leverages both online and in-store shopping behavior data to optimize a personalized shopping experience.
►Do customers readily share their data?
If shoppers have a good experience, and If the retailer continues to provide content of value and relevance, shoppers are ordinarily happy to share personal information. Bonobos, one of our clients, does a great job doing that. Each time I come, they recognize me, they know what I like, what I bought in the past, what fit me, they even offer me a glass of champagne while showing me products I will like. I know I am paying a premium for that, but I don’t mind. The more I have a good experience in store, the more I will come back.
►What is your second announcement?
We believe the future in retail will see smarter machines doing more of the work that is best left to machines, freeing up humans to do the work they’re best suited for, like tending to the needs of shoppers. For example, there is no reason to have humans in the process of finding out there will be an out of stock situation with inventories. Sale associates are there to help me, as a consumer, find the product I want and deliver on my needs. So, in an effort to help staff become more shopper-focused, we’ve built the Performance dashboard to increase efficiency of the store management with predictive and actionable insights.
With the Performance Dashboard, store personnel can quickly gauge performance with goals and the performance of peer stores, both inside and outside of the company, and best of all, the displays a prescriptive set of recommendations to improve store performance, built through a combination of historical data and AI-generated forecasts.
- It acts on AI-predicted traffic: using multiple data sources, including historical traffic and weather data, It predicts what will be the traffic today, by hour, uncovering peak hours and opportunities so the store manager can organize, for example, breaks and lunch times accordingly for his staff. We see often two kind of errors in a store: too many sale associates or not enough. The dashboard knows the key elements to focus today to perform to your goals, and there is a real benefit for both shoppers and the store.
- It identifies key performance metrics – sales, traffic, conversion and ATV – within seconds by benchmarking performance against both goals and peers. For example, as a store manager, it tells me that my store is doing 2% better than the store just like mine, or on the contrary, it might tell me what store is doing better than mine and help me determine how to raise my performance. Instead of relying on intuition or spending time manually analyzing data, the Performance Dashboard tells store managers what to do, and allows them to focus on increasing productivity and engaging with staff and customers.
- If your store is doing poorly, it gives you recommendations for remedies to try, so you know the key elements to focus on today to better perform.
►Can you give us an example?
A leading global sunglasses brand struggled to understand why sales in its signature store in New York were dropping. Looking at the RetailNext Performance Dashboard, it was recommended to contact the top performing San Francisco store for operational tips. The New York store manager quickly realized she had missed the mark on correctly advertising in-store promotions. These actionable recommendations enabled her to adjust her marketing activations, resulting in driving 6-7 percent increase in sales.
►You also announce a collaboration with Facebook
Retailers invest a lot of money in Facebook campaigns and they wonder what the most effective campaigns are in driving traffic in their stores, as well as the demographics of their shoppers, in order to refine and better target their Facebook’ campaigns.
As nearly everybody entering a store has the Facebook app, we use Facebook anonymized data with the use of sensors and/or beacons, to know that FB members are in the store, empowering retailers to know very quickly the ROI of their FB campaigns and to develop actionable insights for delivering a better shopper service through their staffing models, store layout, product merchandising and marketing decisions. For this collaboration, Facebook has ensured user-level data will not be compromised as only aggregated and anonymized data will be shared.
Audience demographic insights from Facebook layers considerable texture to RetailNext’s in-store shopper data, and all this information is also in the dashboard.
►Are the shoppers aware the store they visit use their localization data?
How many apps do you have that are location-based? And again, the data from Facebook are anonymized, and the person can always block the Facebook localization functionality in his app.
► Who are your clients?
Nearly 400 retailers in over 75 countries have adopted RetailNext’s analytics software, including Warby Parker, UNTUCKit, Bose, b8ta and more, and RetailNext adds approximately 10 new retailing brands and installs 800 new store locations each and every month.
Case Study – Marine Layer
Marine Layer is a San Francisco-based company with a mission to make the softest, most comfortable t-shirts around. 33+ stores
- Problem solve – It was difficult for Marine Layer to know how each store was doing at the depth and detail needed to run the store operations effectively.
- Solution – Utilizing traffic counts and integrating POS data, each store now sees how the whole company is doing in addition to their own store. They can see each day, the traffic per hour breakdown, and daily comparisons to quickly identify and address opportunities, saving a lot of time for the Director of Store Operation and the Store Manager.
- Results – A single store lift in conversion of 6 points in just one month, after quickly identify and address opportunities. See the video: Mason James, Director of Store Operations at Marine Layer.
- Arun Nair, CTO and co-founder of RetailNext
- Ressources (Case studies and testimonials, Videos, …) and the blog.
- Follow RetailNext on Twitter
Customer Insight repère les innovations en Digital, Mobile et Retail aux Etats-Unis, et aide les entreprises françaises à transposer avec succès ces stratégies ayant fait leur preuve aux U.S. Notre méthode, nos prestations sur-mesure. Parlons ensemble sur Linkedin, par mail ou au 06 12 83 39 28.