Use Case in Depth—IoE Solutions for the Retail Industry
Connecting people, process, data, and things shines the most in the retail industry. In retail, customers are willing to share a lot of data about themselves to get the deal. For instance, they are happy to share information about their shopping preferences, age, gender, household income range, and more on a customer loyalty application to get points or discounts. They are also willing to let retailers capture that information directly from their behavior and their mobile devices during the shopping experience (within the store or online) using Wi-Fi access and other means. Retailers have used loyalty programs for decades to judge customer retention and the effectiveness of their marketing campaigns, among other things.
Similar to the manufacturing example, there is significant power in converting customer data into knowledge. With adequate knowledge about customers, retailers are able to customize and personalize their relationship with the customer.
The following few sections discuss a variety of uses for IoE technologies or frameworks in the retail space, and how they are making a difference for a good number of retailers, and how some of these technologies are improving employee productivity, improving customer experience, and reducing cost.
Queue Management and IoE (Front-Line Checkout Process)
This is a very interesting field of study for retailers and data analytics providers. Queue management is extremely important for all retailers. Significant research has been done about customer patience. Patience of customers vary depending on the queue processing or progress movement. For example, if the queue is progressing, then customers are willing to spend longer time in it. But they are willing to spend a lot less time in a queue that is not progressing or not moving. Retailers, banks, and even hospitals (i.e., emergency room) have always kept an eye on queues lengths and processing time while coming up with various ways to manage them.
In a retail setting and around the area of frontline checkout, research have shown that customers are not willing to wait for more than two to three minutes in a queue that is seeing little to no progress. But they will stay in a queue that is progressing for close to six minutes. Of course, this does not apply to the queue where the latest Apple iPhone is sold, where the wait-times were several hours.
Imagine a hypothetical scenario of a famous bakery or bagel shop that has 500 stores where an average customer spends five dollars between the hours of 7:00 a.m. and 10:00 a.m. The bakery was noticing that customers were abandoning the queue in almost all of their stores quite often. There are many ways to conduct the queue management studies. Queue management has been the subject of research for a few decades now. This is how banks and retailers decide on how many teller stations or frontline checkout registers to open at any time of day. The studies are mostly manual and require capturing data over a long period of time to improve accuracy.
With IoE-enabled technologies, we have new ways of performing real-time queue management, detecting abandonment rates, comparing them with predetermined thresholds, and alerting store managers to the need for additional staff to enhance the customer experience.
The bakery decides to use an IoE solution where additional cameras are installed to monitor the queue and to capture how many customers are in the queue at any time and how many customers abandon the queue and after how long of a wait-time. The solution is simple yet very powerful. The queue cameras (or possibly new models of existing surveillance cameras) give every customer a tag (or ID) and a time stamp and monitor their progress within the queue, when they get serviced, or long before they abandoned the queue.
They determined that customers were abandoning the queue at a rate of three per hour when there are around 10 people or more in the queue. Doing simple math:
3 per-hour × 3 hours per day (7:00 to 10:00 a.m.) × $5 per customer × 500 Stores = $22,500 per day
Multiply that by 300 days (removing Sundays and few holidays from the calculations) and we get $6.75 million of lost revenue.
Beyond studying the abandonment rate, the system is subsequently programmed to alert the store management in real time to the fact that customers have been in the queue long enough to start seeing abandonment and that additional allocation of staff is needed at the service counter.
Figure 10.1 shows how the camera and the queue-management system could color every customer based on how long they have been in the queue. For example, dotted line “…..” could be for customers who would be serviced in the next two to three minutes, the dashes “------” for customers in the queue close to six minutes, and the solid line “____” for customers expected to be in the queue for more than seven minutes.
Figure 10.1 Queue abandonment scenario
There are endless possibilities for using the queue data. This data could be used across all the stores, or across stores in a specific region to decide on staffing needs, coupons, and incentives, as well as lead to additional process and operational evaluations, to speed up the processing of customers.
The hypothetical example used earlier was only to demonstrate the value of data collected about the customers’ time in the store and connecting the data with other applications and systems to suggest to the store managers ways for addressing the issues. For example, in a large retailer setting, where there are multiple queues and queue types, the calculations and thresholds may be different, but the idea is the same.
IoE Solution for “On-Shelf Availability”
As we mentioned earlier, the customer experience is important in maintaining loyalty. A big part of loyalty is not just having the product in stock but also on the shelf when the customer is looking for it. Today, the process of keeping the shelves stocked is a manual process that is based on time of day or based on visual inspection of shelves. The manual process has gaps and does not always address the issue quick enough to reduce lost sales. The use of point of sale (POS) data is a viable measurement method for many store formats. There are a number of companies that have developed algorithms to estimate out-of-stock from POS data, and some retailers have developed their own in-house systems. The accuracy of estimating out-of-stock using POS data is 85 percent or greater,* which is equivalent or greater to the accuracy of manual audits (where human error is present).
Customer’s reaction to out-of-stock varies from buying an alternative product, not buying an alternative product, to buying at a different store. Approximately, 31 percent of customers buy the product at a different store (Guren). On the average, retailers lose four percent of their annual sales due to out-of-stock issues.† This often translates into long term customer loyalty issues.
Our definition of out-of-stock here is focused on what the customer experiences. It does not matter if we have the product in the back-room. What matters is that the product is not on the shelf. Various solutions (that complement POS-based solutions) have been tested and implemented in the retail space to combat this issue. Sensor-based solutions that alert the store clerks or employees to the missing product have been proposed and in some cases have been implemented. In this scenario, the sensor could be an RFID sensor on the merchandise box itself or a shelf-based sensor measuring the availability and the count of the product on the shelf. But that system is not fully accurate and cannot completely solve the issue. A misplaced product is not easily detected by the RFID scanner if it is outside the range of detection. The shelf capacity scanner will report availability if anything is on the shelf regardless of the correct product or not.
The IoE approach to get the highest accuracy on-shelf-availability measurement is to combine various cost-effective methods and correlate the data together to ensure success. For example, with the advancement of video analytics, we will be able to recognize the product in question by focusing the camera directly on it and analyzing the quantity and placement (ensure that the right product is in the right place). We have seen few solutions where retailers use shelf-based sensors to monitor shelf availability. They quickly found out that it was not a very good solution when their sales associates were putting the wrong products on the wrong shelf. When the sensors were reporting availability, customers were complaining about not finding their favorite cereal box.
A good IoE solution will correlate multiple sources of data to ensure a near-perfect on-shelf availability. A solution that uses shelf-sensors, an advanced video-analytics system that determines type, quantity, and positioning of product, combined with data from the POS system to determine availability of the product on the shelf or in the storage room. If the product is in the storage room, then a real-time alert is generated and sent to the store clerk, manager, or associate to restock the shelf.
Augmented Reality
Augmented reality is gaining a lot of momentum in retail. We believe it is going to be an important part of the customer experience in the very near future. It is already making its way into various major retailers and is making an impact. The concept is not new. More than 12 years ago, hair designers utilized software applications that would superimpose a multitude of hair styles and colors on a picture of a client. Clients loved it. It is not hard to imagine the potential of augmented reality using today’s hardware and software technologies. Look at the following excerpt from an IKEA press release:
Catalog App for smartphones and tablets (iOS and Android)— Download from your app store beginning on July 24. The app gives users access to extended catalog content by scanning designated pages of the printed catalog. The extended content includes: an augmented reality “Place in Your Room” feature which allows users to virtually place and view nearly 300 IKEA products in their own homes; shareable videos featuring quick DIY tips and stories behind IKEA products; 360º views that allow users to look all the way around a whole room; and image galleries. Select content will also be available in the digital catalog.‡
Using this app, you can visualize (or virtually position) how various pieces of furniture from the catalog look in your living room or office by simply pointing at the space where the furniture is needed.
Let us take it a little further: You go to the shoe department of your favorite department store and the sales associate is busy or not available. You see a pair of shoes you like and you need to know the price, availability, color choices, country of manufacture or origin, brand, sales, coupons, and so on. You pick your smart device, you point it directly at the pair of shoes you like, and you have all your questions answered. You like what you see and you want to buy it but the desired size is not available, no problem, the app will tell you if it is available at another store nearby and will ask you if you want it be put on-hold for you. Instead, you prefer to buy it from the app and have it shipped directly to your home.
Similar to all the examples we listed before, this is a true IoE enabled service. We will say it one more time, people, process, data, and things are exactly what went into making the services from the given example possible.
Figure 10.2 is a simplified view of data-flow resulting from interaction among multiple technologies, applications, and things. For example, a customer enters the store, his/her smart device associates with the store’s Wi-Fi system, the customer then uses the store-specific App on a smart device for various reasons (e.g. coupons, deals, latest fashions), the App relays information to the backend applications about the customer and his/her most recent interests and product-searches. At this moment, the retailer knows few things about the customer and is able to “customize” the shopping experience. The customer now walks around the store and sees a pair of shoes that he/she likes and would like to find out the price, available colors, sizes, customer reviews, and few other things. The customer picks up the phone, points it at the shoes and immediately sees all the desired information displayed on the phone’s screen. In addition to the information, the retailer displays few options like alerting a store clerk to bring out the appropriate size to be tried out, put the shoes on hold at another store since they don’t exist at this store, buy using stored credit-card information and ship to the customer’s home, establish an “alert” to email the customer when price-reductions occur, and many more options that retailer could display to give the customer a memorable shopping experience. The above is a result of various technologies and applications exchanging and correlating information to produce the information the customer is looking for. Wireless and “Location Services” technologies to identify and produce a “location” of the customer within the store or in close proximity of the merchandise, databases for historical customer purchases and interests, Inventory systems to gather data about the desired product (color, size, ...etc) and its whereabouts (e.g. local store, nearby store, outside an acceptable driving radius), as well as payment systems to facilitate the purchase.
Figure 10.2 Augmented reality in retail
In summary, augmented reality is helping a good number of retailers reach new or existing customer segments, and offer them new shopping experiences. Augmented reality dressing rooms are showing up at many high-end retailers and are being accepted by customers as a fun experience and as an alternative to physically trying on garments. Welcome to the world of IoE.
Dwell Profile (Time, Video Analytics, and Path Analysis)
We talked earlier about getting customers to the store. Now they are here, how much time are they spending at the store? And where are they spending it? Dwell time is the time customers are spending at the store. Multiple market research studies have shown that the more time customers spend at the store, the more money they will spend. Therefore, keeping the customer in the store and keeping them interested is important for the bottom line.
We originally thought of writing two independent sections, one about dwell time and one about path analysis. But, we decided to combine the two into what we call “Dwell Profile” (time at the store and the path taken inside the store). From a research point of view, the two may be separated, but from an IoE use-case perspective, we wanted to mention them together and make a case for measuring dwell time, path, heat-maps, and the use of smart-devices and apps to personalize the shopping experience and improve loyalty.
Like the queue management, video analytics play a big role in measuring dwell time and performing path analysis, as well as heat map hot spots, in the store. With the advancement of video and surveillance technologies, retailers are able to get automated detection of customers (even their faces) and track their movement, behavior, and the time they spent at different areas of the store. Same devices and technologies will be able to provide heat maps of the store and allow retailers to identify dead-zones or hotspots within the store. This information is very valuable for measuring the effectiveness of displays and other in-store marketing and advertising strategies. We are currently working with retail clients to help them with the dwell time, path taken, and gender of customers that actually visit their displays of big-ticket items. Knowing this information will allow the retailer to position shelf-displays and digital-displays for gender-specific items along the path frequently traveled by customers and also measure the return on investment (ROI) for those displays.
Going back to dwell time for a moment, how could the retailer get the customer to spend more time at the store? Retailers keep their customers occupied and roaming the stores using a number of conventional methods that include entertainment, live music, and free stuff (free samples of merchandise and free tastings of food). What additional ways can the world of IoE offer the retailer in this department? There are two key technologies that can help here. The smart-devices, smart-displays (digital signage—the use of interactive displays, in other words, displays with interactive content) provide customers with additional product and promotional information and allows them to find the product and purchase in a self-service fashion.
Figure 10.3 Dwell-time and path analysis example
In big-box stores (warehouse-like speciality stores or super-centers offering multiple categories of merchandise), dwell-time and path analysis will offer significant advantages. Figure 10.3 depicts a large super store where multiple technologies collaborate to offer the retailer few pieces of information about their store setup and about their customers. The retailer is using Wi-Fi, cameras, and possibly other data-points to profile a particular customer or to measure traffic within specific areas of the store. It also depicts, as an example, two different customers coming through different doors and one having a specific target while the other one roaming the stores through various aisles. Another thing to notice here is the Pan-Tilt, and Zoom (PTZ) cameras at the perimeter used for multiple purposes like keeping track of a customer or zoom on a particular product or shelf to held determine On-Shelf Availability described in an earlier section.
Smart devices will also play a role in increasing dwell time. Research shows that 83 percent of adults aged 18 to 29 and 74 percent of adults aged 30 to 49 have smart phones. In all, 67 percent of cell-phone owners find themselves checking their phone for messages, alerts, or calls even when they don’t notice their phone ringing or vibrating.§ Loyal customers with store-related-apps can get the latest promotions and discounts in real time. But wait, what about the concept of me-tailing? What if a retailer can customize or personalize the shopping experience? Now that I have determined the identity of the customer, having found a few historical facts about him or her, and knowing what isles they have been to and around what sections they lingered, how about pushing them few personalized coupons to help them make a decision in my favor: spending money. This customized experience is the result of many data sources working together to help the retailer profile the customer’s interests in real time based on historical and current information.
Digital Wallets
We could not talk about retail and not mention digital wallets. There are many digital wallet systems, applications, and devices out there. Whether they are web-based, smartphone app-based, or NFC (near field communication) based, they serve the same purpose of making it easier for customers to conduct financial transactions. For our purposes, we want to point out that digital wallets tell a lot about their owners and their habits while also offering a vehicle for loyalty programs for a number of retailers. Digital wallets are a great source of information for payment-system providers and retailers.