Tuning video tracking to the digital customer

    This is a guest post for Internet of Business by Cliff Crosbie in his capacity as SVP for global retail at Prism — the firm’s cloud service transforms any video camera into a business intelligence tool that can be accessed from any device. Retailers large and small use Prism’s platform to remotely audit, manage and optimize their real-world businesses.

    Architecting and operating a modern retail business is tough; companies today will seek out an almost infinite variety of data analytics, customer surveys and market monitoring tools. Prism’s Crosbie argues that sample data sets work fine when it comes to polls, but they are practically useless for understanding a real world retail environment. The very nature of retail is changing… Crosbie writes as follows.

    The complete picture

    Retailers who have to continually make decisions about promotions, merchandising, store layout and staffing need a complete picture of what’s happening in their store: how many people are coming through the door, at different times of day and week; which promotions are attracting the most traffic; which items are being handled the most.

    Crucially though, it’s knowing the shape of journey that customers take through a store that adds so much valuable insight.

    For example, if several visitors walk through the door throughout the day, make a beeline to a specific shoe, pick it up and then put it back on the shelf without buying, this should be a red flag. The shoe is likely to be either too expensive, an unpopular size or damaged. Having the full information to quickly diagnose problems like this and turn around lost sales is pure gold to a store manager.

    A mobile disconnect

    There is no way this full data picture can reliably be collected from tracking users’ mobile phones where consent is given.

    The major pitfall with mobile data gathering technology is that for it to be a reliable source of data, shoppers first have to download and register on an app on a smartphone, AND opt in, AND have Bluetooth turned on. That’s a lot of gates. Shoppers are increasingly reluctant to pass through all these gates due to privacy concerns, not wanting to be bombarded with marketing material or not wanting Bluetooth to drain their phones’ batteries. Or frankly, just not being bothered.

    Shoppers often don’t want to log onto in-store and in-mall WiFi networks for the same reasons. Various technical limitations make WiFi a poor choice for people counting and location tracking anyway. To transmit location, shoppers need continuous WiFi signal access, which can’t be guaranteed and is severely distorted by people, electronics and other moving objects. At best, using WiFi to triangulate a shopper’s position only gives a rough approximation.

    To compete effectively, retailers need in-store technology that provides all the benefits of e-commerce – verifiable counting accuracy, full customer paths, access to real-time counting data and on-demand reports, quantified traffic and dwell data from any store display area, and privacy protection. It’s a long list of requirements, but fortunately recent advances in computer vision have made it possible to meet them.

    The case for IoT analytics

    Unlike mobile data gathering and WiFi, video cameras that have been set up as IoT sensors can count and follow the path of every store visitor. Camera accuracy doesn’t rely on triangulation – determining the location of a point by forming triangles to it from known points. Each element in the camera’s field of view has its precise location and path movements captured in the visual grid. A modern video analytics system can tell a store manager everything they need to know about store visitors and the paths they took on their journeys. It’s exactly the same information that online shops gather, but in a physical environment.

    You may be concerned that a video-based data gathering system would consume so much bandwidth and storage as to be impractical. However, the latest video compression technology is extremely efficient, reducing bandwidth to 1/100th the size of HD video, while keeping the information that’s important crystal clear. Retail staff can access the video footage on any 3G, 4G, or Wi-Fi device – a far cry from what’s possible with popular location-based technologies.

    Having it all 

    In the wake of the Snowden affair, consumers became deeply concerned about privacy, prompting the EU to tighten data laws. This heightened awareness is prompting many consumers to reject consumer apps, especially those with location tracking, and opt out of marketing schemes. Security flaws have also been found in Bluetooth and WiFi tracking devices, adding to this consumer concern.

    Surprisingly, the latest advancements in video IoT make it possible for retailers to gather the most useful aggregate data about store visitors, while protecting individual’s privacy – which may sound very counterintuitive. Video IoT does not require shoppers to register on an app or supply any personal details. New video technology is available that senses people’s presence and movement but strips out their identities from video footage. This unique approach to privacy makes the technology compliant with Safe Harbour laws and tighter EU regulations. Crucially, however, in the event of an armed robbery or other crime, security staff members with appropriate access levels can revisit the footage through and ‘enhanced lens’ that put people back into view so criminals can be tracked down.

    So if you’re a retailer who wants it all – valuable data to improve profits and the in-store experience, consumer privacy and security – the best way forward today is Video IoT.