Home → Magazine Archive → February 2020 (Vol. 63, No. 2) → Tracking Shoppers → Full Text

Tracking Shoppers

By Keith Kirkpatrick

Communications of the ACM, Vol. 63 No. 2, Pages 19-21
10.1145/3374876

[article image]

Save PDF

Most people are aware that if they use the Internet, social media sites, or electronic payment systems, their activity is tracked, thereby creating a digital footprint and roadmap that can reveal a significant amount of personal data and activity patterns. Online retailers and marketers are leveraging (some may say exploiting) this treasure trove of data, helping them craft advertisements and marketing messages that specifically target consumers in the hopes of driving sales, or creating a deeper level of engagement with customers.

It is no surprise physical brick-and-mortar retailers are trying to replicate the engagement strategies that have worked well for their online counterparts (or which they themselves have used online). Using a combination of technology and social engineering, retailers are tracking shoppers in and around their stores to leverage shopping and behavioral data to increase sales, build greater customer loyalty and, in some cases, deter or thwart theft or fraud.

Indeed, according to Retail Systems Research, a research company that tracks technology use by predominantly North American retailers, 61% of those surveyed in 2018 said they see potential value in technology that tracks shoppers as they walk through the store, up from 51% the year before. Similarly, the use of in-store cameras and video analytics for customer experience purposes hit an interest level of 49% in 2018, up from 43% in 2017.

Not surprisingly, the use of video surveillance technology appears to be more pervasive in other parts of the world, such as Asia, where consumer privacy protections are weaker and shoppers appear to be more receptive to the use of new technology, provided they receive tangible benefits.

uf1.jpg
Figure. Thermal after-images of shoppers can be tracked as a "heat map," to inform retailers of shopper traffic patterns in their stores.

"Especially in the APAC (Asia-Pacific) region, you see a lot of companies that will use surveillance footage that will capture the images of customers, and then match that to their loyalty programs," explains Auria Moore, director of solution strategy at Stibo Systems, a provider of master data management (MDM) solutions that allow retailers to maintain a single view of all customer data. "Some companies are doing it better than others."

Cosmetics seller Sephora is an example of a retailer that has successfully deployed video surveillance and analytics to provide a more granular view of store traffic patterns, while also providing more individualized marketing and incentives to customers. Working with AllGoVision, a provider of video analytics solutions, Sephora deployed 54 Axis cameras at store entrances and near various product sections across 10 Sephora retail outlets located in Malaysia.

The video solution can count the number of people entering and moving about the store, and feeds this data into the analytics platform. The data is used to monitor and address checkout line wait times, and to generate store traffic flow maps and heat maps; the data is then fed into a model that can be used to optimize the retail space. The solution also incorporates facial detection and recognition technology, which allows Sephora to better measure product engagement with individual users, create demographic and marketing profiles for products, and can be used to aid in shoplifter detection and prevention.

Camera data that includes customer positioning, daypart engagement information, and foot traffic can be combined with data analytics "to help marketing make decisions about how you place a promotion, what time of day you have more people passing by, and how you set up your retail space, all based on the data you've collected," says Aji Anirudhan, global sales and marketing manager for AllGoVision, Inc.

However, the use of cameras feels invasive to some shoppers, which is why some retailers choose a different method of tracking shoppers. Beacons—small devices that send signals to Bluetooth-equipped devices such as smartphones—can be used to detect the presence of shoppers, recognize them individually, and interact with each one by providing an offer (such as a coupon) to further strengthen the store/customer relationship.

Retail chains with a presence in the U.S., such as Target, Walmart, Urban Outfitters, and Nieman Marcus, have deployed such Bluetooth-based wireless technology to connect and interact with customers over the past few years. Unlike Wi-Fi, beacons work well indoors, can track customer location within stores with accuracy ranging from within a foot to a few yards, and are energy-efficient, making them ideal for the retail environment.

Target, one of the notable large retailers to embrace beacons a few years ago, had rolled out the technology to most of its stores by last year. Shoppers who download the Target app (which is equipped with software that interacts with the beacons) will display their location on the app's map as they move throughout the store.

Mike McNamara, Target's CIO, explained in a video discussing the technology that the store-tracking technology is "a bit like driving your car with GPS. Just click on an item from your list, and the app will indicate on a store map the precise aisle where you will find your item." McNamara noted that the app will even tell you if the product is on sale.

Because beacons are designed to tie offline behavior (shopping in a physical store) to online behavior (which is often tracked by the use of a store app or via browsing history on a mobile or desktop device), retailers can capture, analyze, and make predictions on a substantial amount of personal data. A retailer then can use this data to increase both online and offline sales by offering discounts, loyalty points, or other incentives in order to further engage customers and drive sales. That means consumers' private habits are quickly becoming public knowledge, which is then used for commercial purposes.

"I still see that people are willing to trade the benefit of giving up their personal information," says Norman Shaw, an associate professor at Ryerson University in Toronto, Canada, who focuses on digital privacy issues. "They want to give it up because they get the convenience."

In a retail environment, a beacon placed adjacent to a product would be alerted to a shopper's proximity to the beacon by the app on the shopper's phone. Once the smartphone app recognizes the beacon, it sends the data back to a company's server, and the app then can be prompted to display ads for products the customer is likely to buy, based on data contained in that shopper's profile. Alternatively, a store can use a strategy called "retargeting," which refers to the practice of sending a follow-up offer, such as a coupon, after the shopper leaves the store, to entice the shopper to return.


The tide may be changing, as there appears to be growing interest in the use of beacons among retailers.


Despite their potential benefits to both retailers and customers seeking deals or personalized offers, the adoption rate of beacons has not matched the hype. Indeed, much of the coverage of beacons being deployed at large retailers such as Walmart, Target, and other retailers occurred in the 2013–2017 timeframe; since then, few success stories have been popping up in the popular or trade press. "Beacons are valuable, [but] they blend the lines of scariness at the same time," Moore says. "I don't think retailers are adopting them as quickly as they thought."

This could be due to growing consumer concerns about privacy. Most large, reputable retailers notify customers they may be tracked if they download their mobile application, although these disclosures are often buried within their terms and conditions, rather than being highlighted within the application itself. However, a large number of third-party location-marketing firms generate beacon tracking code and bundle it into a tool kit that developers can insert into benign-looking weather or newsfeed apps. A retailer would then be able to track where a customer spends the most time in a supermarket, even if that user wasn't aware he or she was being explicitly tracked.

That said, the tide may be changing, as there seems to be increasing interest in beacons among retailers, according to data from Retail Systems Research. The use of beacons for in-store marketing applications saw an 11% year-over-year increase in interest among retailers between 2017 and 2018, from 38% to 49%. Similarly, the use of beacons for store perimeter marketing saw an interest level of 44% in 2018, up from 37% in 2017.

"The interest [in beacons]" is there, Rowen says, noting that he expects this year interest in beacons for in-store marketing is likely to reach nearly 60% of all U.S. retailers, based on the established historical trajectory of retailer interest in beacons. From the retailers' perspective, the biggest challenge is that there are so many potential technological innovations that could be used, that picking a winner or devising a strategy has become a massive challenge.

Indeed, Rowen says many retailers have been afflicted with the 'if I buy something this year, who's to say it's still going to be valuable next year?' conundrum, particularly because store margins are so thin that any mistake likely will have a huge, potentially devastating financial impact.

"There are these exciting technologies that are available," Rowen says. "But if [they're] not guaranteed that they will increase sales, reduce shrink (theft), increase footfall (the number of people visiting a store), drive conversion (from shopping to buying), or prompt ancillary sales," stores are unlikely to commit operational expenditures to deploying a new technological solution.

Tracking shoppers, however, is not solely about marketing. Retailers are also keen to use technology to track shoppers to identify potential shoplifters. Shrinkage, in industry parlance, remains a major problem for retailers, and due to the legal and safety issues surrounding the apprehension of thieves after they have taken merchandise or cash, there is a push to identify and thwart theft prior to it occurring.

A combination of video surveillance and machine learning can be used to not only identify people who populate the BOLO ("be on the lookout for") lists of suspected or known thieves, but also those who may display body-language cues (nervous tics, eyes scanning the area rapidly, etc.) that could indicate they may be looking to steal.

Vendors including FaceFirst, Vaak, and Third Eye have conducted extensive public relations campaigns highlighting the effectiveness of their solutions; FaceFirst, for example, claimed its software is in use at "hundreds" of retail locations. However, it is difficult to accurately quantify what percentage of retailers are actively using facial recognition or video analytics to deter or prevent fraud and theft, because few retailers choose to publicize their use of the technology, lest they incur the ire and scorn of privacy-minded shoppers.

Anirudhan says retailers outside the U.S. may be somewhat more comfortable discussing how they use video technology, possibly because privacy is not expected in some markets, such as Asia. Using video analytics for marketing purposes, such as making the store easier to navigate, or by offering shoppers discounts or other benefits, is likely to be acceptable to shoppers in those regions, but using the technology to deter theft may not be as welcome as there is no direct benefit to shoppers.

Further, facial recognition technology, as well as gesture or body-language analysis technologies, are not yet accurate or reliable enough for some retailers to use, mainly because there isn't a large-enough repository of real-world video data available that can be used to train recognition systems so they are accurate. Particularly in privacy-conscious markets such as the U.S., these types of systems need to be accurate nearly 100% of the time in order to gain the confidence of retailers.

Currently, regulation of video data varies by jurisdiction. In more autocratic jurisdictions, such as China, it's unlikely that meaningful regulation on the use of video will be enacted. However, in privacy-focused Europe, the General Data Protection Regulation (GDPR) appears to apply to not just text-based data, but also video. In October 2018, the Austrian regulator fined a company 4,800 euros for capturing detailed images of passers-by and not having adequate signage to alert them their images were being captured.

In Canada, the Information and Privacy Commissioner of Ontario released in 2015 updated guidelines governing the lawful use of video surveillance, which laid out a number of collection, use, access, and retention requirements.

While there is no law in the U.S. currently requiring retailers to disclose their customers are being tracked or recorded on video, that may change, according to Moore.

"I think you're going to see digital compliance tightness spill over into the retail store, and I don't know what that looks like yet," Moore says. "It could be as simple as a sign that hangs in front of every Walmart store that says, 'video surveillance is in use and you agree to be captured for [marketing or security purposes].' But in order to get shoppers to provide this consent, retailers will need to provide a worthwhile trade-off for the use of shoppers' private data," Moore says.

"My data is my gold, so if I'm giving up my gold, what am I getting back?" Moore says. "I think that the experience has to match up for consumers. You look at things like the Fitbit and Apple Watch, people are okay with giving up their data if it's going to help them monitor their heart. You're okay with giving up data if it's going to help them in their day-to-day lives. [Giving up privacy] for a coupon or promo code while I'm in the store is not good enough."

* Further Reading

Kilcourse, B. and Rowen, S.
Location Intelligence in Retail: The Value of 'Where', Retail Systems Research, February 27, 2019 https://www.rsrresearch.com/research/location-intelligence-in-retail-the-value-of-where

Lang, E.
Store People Counting and Gesture Recognition, December 20, 2018 https://www.youtube.com/watch?v=cryEd6XD41k

There's More in Store With the Target App, A Bullseye View, September 20, 2017 https://corporate.target.com/article/2017/09/target-app-mike-mcnamara

Back to Top

Author

Keith Kirkpatrick is principal of 4K Research & Consulting, LLC, based in New York.


©2020 ACM  0001-0782/20/2

Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or fee. Request permission to publish from [email protected] or fax (212) 869-0481.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2020 ACM, Inc.

0 Comments

No entries found