It’s been a while since my last post, busy days. I am back with snack posts, a new format that takes only a few minutes to read.
Intelligent retail (aka smart retail) is enabled by technologies such as Internet of Things (IoT), Big data, data analytics and Artificial Intelligence (AI). It helps retailers to make better business decisions and to provide a seamless customer experience. Business decisions such as: what is the most effective place in the store to locate a new promotion? Or what is the level of staff needed in the next quarter to drive sales?.
The state-of-the-art technologies enable solutions that can be implemented today in any fashion store, supermarket or department store to help retailers make more effective business decisions based on data. And how does it work? i) IoT connected devices (sensors) are deployed in the store, measuring customer activity as footfall (visits), shopper frequency and recency, dwell time… ii) the new data from sensors is processed and combined, and data analytics generates powerful insight iii) the retailer makes data-driven decisions based on the insight provided (visualised on a dashboard or elaborated on a business report).
But (there is always a but in these articles) despite technology enables it, the business case doesn’t work for most retailers implementing a complete IoT retail proposition. Let me explain the reasons why and make some suggestions to fix it.
[ordering the snack]: Intelligent Retail solves retailers’ pain points
Retailers pain points are commonly related to two critical business performance drivers: increase sales, and save cost by improving operational efficiency.
Some years ago, I led the development of an innovative product to help retailers engage customers in understanding their behaviour in-store, like the “google analytics for the physical stores”. My team and I had the opportunity to discuss with retailers and customers getting a real understanding of the problem, still valid nowadays.
Regarding indoor needs, retailers identified and mixed the following topics across the customer journey (pre-purchase-post): i) to identify customer behaviour in-store (indoor insight) ii) to turn insight into sales and loyalty (engagement) iii) to optimise omnichannel (a multi-channel strategy providing a seamless purchase experience whether the customer is shopping online, from mobile or at brick-and-mortar stores iv) to make payment easier in physical stores and v) to improve operational efficiency.
The truth is that brick-and-mortar operates blindly compared to e-commerce retailers (that have complete shopper behaviour insight). Despite having critical information on sales and visits, they miss data such as new vs repeat customers, how long they stay, where shopper dwell the most, how they move between zones, etc. Intelligent retail bridges that gap.
[serving the snack]: IoT Retail helps retailers to make more effective business decisions on operations, marketing and merchandising
Intelligent retail provides sophisticated insight that benefits each business area. Some examples are:
Operations: IoT provides service level insight to meet customer expectations across the store, for instance, eliminating long queues to reduce the waiting and transaction time, schedule staff level efficiently, detect stock levels and minimise equipment downtime.
Merchandising: IoT data is used to evaluate the effectiveness of store layouts and product interaction. The insight generated allows identifying customer preferences and the path to purchase.
Marketing: measuring the effectiveness of promotions, identify customer preferences to send deals to customers in the store, deliver curated content to cross-sell and up-sell.
Although any solution can be implemented and the retailers recognise that Intelligent retail solves their pains, the main issue is to make the business case works. This problem is more significant when deploying a massive IoT solution: sensors covering the entire store and capturing all the shoppers’ activity and behaviour.
IoT devices (hardware cost) and advanced analytics are the two critical components of the pricing model to prepare a successful business case. The other two pieces, connectivity and the IoT platform (software cost) give more flexibility to make the numbers right. IoT Retail uses a combination of devices such as video cameras, bluetooth beacons, wifi, RFID tags, smart shelves and even mobile networks (small cells) to capture data about the customer behaviour in the store. The elevated price of some of them, especially video cameras, make the business case fails. Advanced analytics uses machine learning and Artificial intelligence to detect patterns faster and make predictions and often cost more than the retailer would want to pay (despite it would bring huge benefits).
[eating the snack]: Take a lean approach to succeed
The good news is that there are successful stories (I got some) when you take a lean approach and follow these tips:
- Forget about deploying sensors across the whole store to get the “google analytics for the physical store” dream. I did that on my first try, it failed and I learned.
- Instead, consider a lean approach, preparing a priority list for the retailer’s need locating sensors just in a few critical zones that match those priorities. When I started doing that, everything worked seamlessly, for example, measuring the effectiveness of a promotion (zones).
- Be innovative in sensors and technology. A practical alternative might be to use existing surveillance cameras instead of deploying only new expensive video cameras across the store. A computer vision software will do the rest (but sharing only metadata and not the real images). A radical approach could be using 4D radar imaging technology to get affordable and highly advanced sensors (replacing most 3D cameras).
- The proposition should be modular, adding complexity and features progressively according to the retailer’s experience with in-store insight.
- Regarding that modular approach, the predictive analytics using machine learning and Artificial intelligence (to forecast sales and footfall traffic) will be at the upper tier, and it would come last. Don’t get me wrong, it is the most important piece but i) historical data need to be recorded at least during six months (and be properly cleaned) to get a useful insight to make business decision ii) the retailer needs to build confidence in data and check that investing in data provides a proper return of investment (ROI).
- Be creative in the business model. As an excellent example, some companies providing predictive analytics, offer a guaranteed model, so retailers don’t pay for wrong forecasts.
I hope you enjoyed the snack!
