Imagine owning a store and knowing everything about your customers from the moment they walk in: ages, genders, moods, information from their social media profiles. It seems like something out of Minority Report.
Except, it's nearly reality. Stores regularly track demographic and biometric data like age, gender, and mood just from reading faces of customers on video cameras. The only reason they can't pull social media information is because networks like Facebook and LinkedIn don't allow that kind of third-party access to their systems.
This data can be invaluable to retailers, who can use it to target their core demographics in real time and sell it to brands or other third parties. It's a large part of why the video analytics market grew to $1.69 billion in 2016 and is expected to surpass $4 billion by 2021, according to a report last year by research firm MarketsandMarkets.
That doesn't necessarily mean it's the best option for small- and medium-sized businesses: Some in the industry are moving away from demographic and biometric data. Whether that's in the name of privacy protection or due to fear of customer backlash, the end result is the same. So how do small retailers decide whether this type of data collection is worth pursuing? It starts with understanding its benefits and drawbacks.
A more personalized future
Some video analytics companies are going all-in on demographic data collection. San Jose, California-based RetailNext, for example, uses video to track consumer age and gender. Countbox, which was founded in Moscow in 2008 and now operates across Europe and North America, uses "demographic sensors"--essentially, video cameras with on-board cloud computing software--to discern store visitors' age, gender, ethnicity, and even mood.
Countbox co-founder and CIO Andrey Stryukov says the technology will become more personalized over the next decade. "It's not a very hard task to match the face of a person who enters the store to a database of faces," Stryukov says. "In the future, who knows? Facebook may ... do something in-house like that or may allow a company like ours to access their faces database."
Indeed, the burden of preventing overstepping lies with private companies like Facebook rather than the federal government. Illinois and Texas have laws against excessive biometric data collection, and other states have proposed similar legislation in recent months, but the Federal Trade Commission--the primary agency that handles federal prosecution of businesses--can't go after companies for it. The FTC can prosecute businesses for deceptive acts and practices, such as if a company offers an opt-out policy but never actually removes those who opt out from the system. (That's exactly what happened to analytics company Nomi Technologies in 2015.) Still, there's an easy way around that: Opt-in and opt-out agreements, while best practices, are not required by federal law.
Analytics companies that neglect to alert their customers or intentionally avoid doing so, however, should be ready for significant backlash. San Francisco-based Euclid Analytics lost customers like Nordstrom in 2013 and coffee chain Philz Coffee in 2014 after consumers expressed concerns over their mobile devices being tracked in stores. Euclid did offer an opt-out in the privacy tab on its website--it still does--but has no obligation to reveal its client list, meaning it's hard for consumers to know if they regularly visit Euclid-tracked locations.
For now, situations like that tend to be the exception. RetailNext spokeswoman Bridget Johns says retailers often outpace analytics companies in this regard. "Retailers do a really good job of staying far ahead of the law," Johns says. "Things that may be legal to do, if a retailer doesn't think that they are delivering a value and are in alignment with their customer expectations, they just won't do it."
Stryukov notes the Philz Coffee example as a wake-up call. "After this case, analytics providers--companies like ours--we're being cautious about how we're collecting data," he says, before adding that consumers will eventually need to face an inevitable future. "It's going to be more accurate customer profiling in the store, but [right now] everyone's afraid of that. No one wants that."
"There's always going to be a market for privacy"
A segment of the video analytics industry believes there's another way. LNL Systems, an Inc. 5000 company in 2015, has developed technology that uses cameras only to track store traffic--data that founder and CEO Mark Barnes argues is more important than demographic and biometric information. Knowing when customers are dwelling in certain areas of stores, he says, helps associates know which consumers need help and which are just browsing. It can also help with loss prevention: If a customer is dwelling in a high-theft area, an associate can get tipped off to check it out.
Privacy is an accidental byproduct of collecting only store traffic data, a business strategy that's cheaper for retailers in the struggling brick-and-mortar retail industry than paying for a full demographics suite . "A retailer might look to deploy both," Barnes says. "It's certainly a lot more expensive to be able to have that software that starts doing the demographics. The majority of retailers now--and I've seen this shift just within the past two or three months--they're worried about 'How do we keep the lights on?'"
Video analytics companies that operate in Europe are more overtly privacy-conscious, especially following last year's European Union regulation banning biometric data collection without a person's "explicit consent." Irisys, a longtime United Kingdom-based company, has recently started using thermal cameras that can't recognize faces. Prism Skylabs, a San Francisco startup operating domestically and internationally, makes software that separates foregrounds and backgrounds--meaning it can erase people from the camera's view altogether if necessary.
"There's always going to be a market for privacy," says Irisys spokesman Jordan Britchford. "It's difficult for anyone who is trying to gather data--on their customers entering a store or whatever--to strike a balance. But there is a balance to be struck there, because if you go too quickly, you're probably going start turning customers off."
Cliff Crosbie, a senior vice president at Prism Skylabs, agrees--though he hopes the technology can eventually evolve toward facial recognition that still protects privacy. "That's probably something that's coming," Crosbie says. "We just need to find a way of doing that."
The data retailers really need
Small businesses certainly shouldn't view the victory of biometric data collection over non-video systems as automatic. Measuring foot traffic, for example, could provide small retailers with stronger immediate takeaways--both inside and outside stores--than extensive demographic breakdowns. "If my neighborhood coffee store is able to sense as I'm walking past it, how could it actually incentivize me to come and grab that sandwich or grab that coffee?" asks Anindya Ghose, a professor at NYU's Stern School of Business and author of Tap: Unlocking the Mobile Economy, a book that focuses partially on the subject. "That would be my message to a lot of the small or medium stores--whether it's through beacons, through Wi-Fi, or other location technologies--that you can incentivize people to come to your store."
Ghose's research finds that as the technology evolves, people may become more receptive to having their biometric data collected for commercial use--but we're still in the early stages of that evolution. "What's basically happening are these structural changes in what is considered socially acceptable by people," he says. "Location data and basic demographic data, people are getting more and more comfortable with. They're not there yet with facial recognition data. My sense, though, is: It's a matter of time."