New Browser Extension Decides How Trustworthy You Are
Source: Jason Tanz
Three years ago, in a TEDGlobal talk, sharing-economy guru Rachel Botsman shared her vision of a “reputation dashboard”―a kind of credit report that tracks your online behavior across services like Airbnb, TaskRabbit, and Dogvacay and compiles it into a portable measurement of your trustworthiness. Amassing that data, Botsman proposed, would make reputation into a kind of currency. “In the 21st century,” she predicted, “new trust networks, and the reputation capital they generate, will reinvent the way we think about wealth, markets, power and personal identity, in ways we can’t yet even imagine.”
It’s a compelling vision, but so far it hasn’t been realized. That’s because, as I noted last year, the companies that have amassed the most reputation data aren’t eager to share it. “We’re in an early and competitive stage,” Monroe Labouisse, Airbnb’s director of customer service, told me at the time. “That asset―the trust, the data, the reputations that people are building―is hugely valuable. So I’m not sure why a company would give that up.”
A new company is trying to do an end-run around that intransigence by scraping publicly available information from various sharing-economy services and compiling it into a trust score between 0 and 100. Called Karma, it works as a browser extension―any time you pull up a supported site (which currently includes Airbnb, Craigslist, Dogvacay, Ebay, Etsy, RelayRides, and Vayable) a pop-up window will ask if you want to link your account to your Karma score. That score is calculated by looking at the reviews you’ve received―both the quantitative ratings (the number of stars, for instance) as well as a textual analysis of written comments. Different services are weighted differently; intimate interactions like those powered by Airbnb and Dogvacay are deemed more relevant than relatively anonymous eBay sales, and more recent reviews also are weighted more heavily. The more services you link, the higher your potential score. (Of course, if you’ve misbehaved on one service, your score could fall―but then, you would probably choose not to link it in the first place.) When you peruse a supported service, you’ll see every user’s Karma score superimposed over their listings. It’s a little bit like the sharing-economy’s answer to Klout, that notorious Q rating for social media.
Zach Schiff-Abrams, Karma’s co-founder and CEO, says the company has not contacted companies like Airbnb or Dogvacay. But he thinks they will welcome his service, because it will make it easier for new hosts to attract guests, instead of grinding through the first few months as they attempt to build up a bank of positive reviews. “TaskRabbit’s biggest frustration point is the on-boarding process for new Rabbits,” Schiff-Abrams says. “We think that Karma can act as an arbiter to help these people begin to build a reputation much sooner.”
There’s something compelling and simple about this. It ignores all the complicated behind-the-scenes algorithms and processes that companies like Airbnb use to establish reputation and just collects the publicly available result of those processes. According to Schiff-Abrams, this simple hack―going directly to the users rather than through the enterprise―is what convinced VCs like Great Oaks to support them.
Gaming the System
These platforms now represent billions of dollars in commerce, and as such they must be extremely wary of bad actors trying to game the system. And Karma’s system, at first blush, looks pretty gameable. Going through the browser makes it easier for Karma to reach users directly, but it also makes it harder to confirm a Karma user’s true identity; if I’m using a friend’s browser, it would be pretty easy to link their Karma score to one of my accounts.
There’s also a weakest-link problem here. It’s easy to imagine using Karma as a Trojan horse―building up a high reputation score on a more easily-gameable system and then importing that score into the fortress of Airbnb.
It’s precisely to avoid that kind of scenario that Airbnb has invested so much money in its trust and safety division, an intricate and detailed set of algorithms to sniff out sketchy behavior. (If a new listing is getting a lot of positive reviews from the same account, for instance, the algorithm will flag it.) Of course those algorithms impact what gets posted on Airbnb, so in some ways Karma is freeloading off Airbnb’s expensive and painstaking security infrastructure.
Right now, Airbnb insures its hosts up to $1 million, in part because it trusts its algorithm to guard against the most egregious forms of fraud or malfeasance. But if someone is using a Karma score to determine who to rent to, that means that Airbnb is suddenly assuming the risk for a different company’s security mechanisms. It’s hard to imagine Airbnb will go for that, and I’d expect them to insert some language saying that anyone who uses Karma is no longer eligible for the insurance coverage―which would probably be enough to do serious damage to Karma. (Airbnb declined to comment on a product they haven’t had a chance to use yet.)
If this were earlier along in the sharing economy, Karma might have a bit more time to work all this out. Certainly that’s what happened to Airbnb―as it grew, it realized it had to get more serious about security, and it had plenty of missteps along the way. But now this is a mature market, and I’m afraid it’s a little late for this kind of experimentation. This is a clever approach to a big problem, and it’s frustrating that competitive pressures are preventing it from getting solved. But ultimately I’m afraid that the problems of identity and trust are too complicated and fraught to be solved with a simple browser extension.
| }
|