You’ve probably heard the term “contact tracing”—perhaps in an argument about whether or not we should let the government use our phones as tracking devices to identify and isolate people infected with COVID-19.
In theory, contact tracing would work in tandem with mass testing to stop the spread of the novel coronavirus. In practice, it raises a host of privacy concerns and other problems that appear when viable technical solutions are developed without considering the behavior of real human users. In short, contact tracing has a UX problem.
Fortunately, it can be corrected.
The rationale behind contact tracing
How does contact tracing work? Traditionally, public health officials track down sick people and sleuth out every single person they came into contact with. It’s a painstaking, manual process. Luckily, it’s now being automated. Earlier this month, Apple and Google announced a collaborative proposal to use our phones’ Bluetooth—a low-power radio normally used to talk to other gadgets, such as wireless headphones—to anonymously record every other phone we’ve been near over time. Our phones already do this to some extent by scanning for other Bluetooth devices, but don’t typically keep a log of those potential connections. Under this proposal, our phones would record each of those hundreds or thousands of little radio hellos in a sort of shared file, which, if we were diagnosed as having the coronavirus, would make it possible to look back through time to see with whom we might have been in contact. You could think of it as the inversion of the way a virus spreads. It wouldn’t tell you where exactly you were exposed to the virus, but if you were to voluntarily report yourself a carrier, your phone would be able to tell every other phone you had been near that you might have exposed its owner to infection.
Collectively, all those phones would be able to start making a map—through both space and time. As if you went to Google Maps and typed in “Where was I on March 3?” and could watch it draw a squiggly line from where you are standing right now, backward to your bed, into your car, yesterday, driving in reverse through town to drop off a bag of groceries at the market, and so on, until the line ended with a dot showing you at your local park watching your child scoot up a slide.
The contact tracing service could then overlay that reversed route with everyone else’s routes. When your route crossed with mine last Thursday, standing 6 feet apart in the line outside our local bakery? That’s a potential contact. Maybe you were a carrier of the virus then, perhaps asymptomatically. I wouldn’t know if you had spread the virus—or if I had spread it to you. But we’d both know it would be a good idea to get tested.
That’s the rationale behind contact tracing. Until we have enough testing kits for everyone on the planet, we need to be able to figure out not just who is sick, but who else they might have gotten sick. Then we can isolate those most at risk, instead of, you know, the more than 8 billion people on the planet.
A good idea that needs better execution
In general, experts—such as this distinguished team of Oxford researchers—think contact tracing is a pretty good idea. And in countries where it has been compulsorily implemented, it has proven to be effective. With fresh memories of a nasty 2015 MERS outbreak, South Korean health authorities are legally permitted to aggregate a smorgasbord of highly revealing data sources once anyone tests positive for COVID-19: GPS location data, credit card payments, CCTV footage, travel documents, and medical records—far beyond the relatively limited tracking suggested by the Apple-Google proposal. With just 10,683 infections as of April 21 (after hitting 7,300 on March 8), the South Koreans have crushed their infection curve. (In contrast, the United States has gone from 450 infections on March 8 to more than 820,000 in that same period.)
Many critics of digital contact tracing locked in immediately on the potential privacy implications of bad actors inside the walls of technology giants, malicious governmental regulators, or even central-casting rogue hackers—fair concerns, to be sure.
And indeed, the South Korean system of alerts—blasting out rather specific personal information about carriers over text message—has already led to some troubling social shaming incidents regarding personal choices people make that have nothing to do with their carrier status. The contact tracing system may be without bias, but it takes little to prompt other people to make biased judgments. “Why were you walking through that dark alley at night?” has nothing to do with a positive or negative exposure to a virus, but don’t try to tell a busybody that.
But there’s another reason contact tracing won’t work in the United States or Europe: It is only effective if more than 60% of the population opts in to the program. That won’t happen as things stand.
The individual versus the collective
The main reason is that contact tracing hasn’t answered the question that any user-friendly product answers: What can it do for the individual? Sure, some people might sign up because it makes them feel good to contribute to the health of the community. But 60%? Singapore, a country with more tolerance for government intrusion than most Western democracies, has only about 16% of the population willing to download the country’s TraceTogether app, as of April 1—roughly 1/5 of what it would take to make the contact tracing effective in the first place.
Instead, I suggest we ask a more pointed question: What would a contact tracing app do for an individual who doesn’t care about the collective society at all? Or to use a slightly out-of-fashion startup coinage: What is the value proposition for the contact tracing user? The only way contact tracing will function at all will be if there is a value to each individual so overwhelmingly obvious that it would outweigh their concerns over privacy—and more realistically, get them to consider participating at all.
I don’t have the answer, although I suspect scaling down and individualizing the user story will be a big part of it. It is hard to convince a person to care about theoretical people represented only by concepts or numbers, but it is easier to remind a person that they care about their friends and neighbors. Identifying the right message to explain why contact tracing would help protect real people that you know and love must surely be part of it. And removing any friction in using contact tracing at all—from before any download to far after an install—is so obvious that it bears repeating again and again.
At my company, a “cybersecurity solutions provider” (or whatever term is in vogue these days), we sometimes talk about how we do not want to be in the “take your medicine” business, where so many security companies seem to traffic. Not because we’re not medicine—ounce of prevention, and all that—but because as any parent knows, simply telling someone to change their behavior “because it’s good for you” isn’t very effective.
But if you can make the medicine part of something more—a game, a treat, a reward—then it goes down much easier. That reward may simply be humanizing the results of your individual action through language: maybe pithy writing or clear illustrations. Make the hassle of taking this small individual risk of downloading an app tangible—”Do It for Kansas City” might resonate better than “Do It for Missouri.” Or flip the passive, opt-in story for one that provokes feelings of unity and togetherness, maybe even with a splash of competition—a live-updating map of the nation showing which states have turned on the most contact tracing apps. Normalize participation and education, à la “I Voted” stickers. Make it a point of pride to show that you’re part of the solution.
Any of those ideas might work—or they might not. But they can be tested, refined, and iterated upon quickly, using user experience and education principles that have been brought to bear in far less important contexts.
Apple and Google have built multibillion-dollar companies on their ability to create experiences that evoke these incentives and emotions—some of the most powerful examples of impactful user experience of the digital age. This isn’t the time to forget those principles, especially when they can result in something more important than any app: saving the lives of our friends and families. Until these companies and governments proposing contact tracing can tell this story from start to finish, from big idea to personal impact, there’s little chance it will be implemented at all. Finish the user story.