Today’s world is one of countless apps and services, but also one of ratings. In the gig economy, everything is rated — and ratings are everything. Ratings boost products and services to the first page of search results, putting them in front of consumers.
Uber, like many companies in the “gig economy,” has been highly scrutinized for its driver ratings service, where the stakes are high. It may seem like a minor point of controversy for the company, which has seen scandal upon scandal — the latest being an FBI investigation for interfering with rivals — but it illustrates the delicate nature of user ratings.
“If your average rating falls below a 4.3 after your first 25 trips,” Uber’s policies state, “Your profile will be deactivated and you will need to take a quality improvement course in order to be considered for reactivation.”
The ride-sharing company’s ratings system has slipped into binary: five stars and everything else. Uber has essentially moved the goalposts so that a four is a zero.
Ratings has always been a tricky game because context is so important. Is a 65% on Rotten Tomatoes good? In academic terms, it’s a D. But the site helps by certifying a film as “fresh,” if over 60% of reviews were positive, and “rotten,” if otherwise.
Unfortunately, Uber reviews aren’t put into context, which can, at times, make the rating system unnecessarily harsh. Some customers may only give a driver five stars when the service has been stellar. However, factors outside of a driver’s control like traffic might unfairly compel a passenger to ding the driver in a rating.
Two years ago, Frances Frei, Uber’s SVP of leadership and strategy and a Harvard Business School professor, called this five-or-bust strategy “close to useless,” arguing that a good management system was key to deal with these problems of uninformed reviewers.
Michael Truong, a senior product manager at Uber, has been trying to fix these issues since he joined the company in 2015. Recently he and his team rolled out some new features to improve the “management” of reviews.
“What we’ve found is that there’s actually a lot of good ratings in the system,” he told Yahoo Finance in an interview. The question then was, “Is this a good thing or bad thing?”
Clearly, it’s a good thing, and it indicates that most Uber drivers fulfill their promise of safely and courteously driving passengers from point A to point B. There is no reason why drivers shouldn’t be able to adhere to such standards.
When reviews are skewed so positively, a company must decide whether it’s working to achieve its goals. YouTube abandoned the five-star method in 2009 after it found that almost all reviews were either five stars or one star. It moved to the thumbs-up system, which is what people said they wanted.
Uber may be in a similar situation in some ways, but Uber sees the five-star system as important. First of all, it’s impossible to get a literal five-star experience without a rating system. But furthermore, a simpler system would remove customers’ ability to give more nuanced feedback.
“We did test a little while back and tried different rating systems to see what they’d do,” Truong said. “The main thing we want to accomplish is seeing if there’s something wrong. A Thumbs-up system is basically: was it good or bad. In these alternate systems you only hear about things when it’s really bad. You don’t only want to know when things are really bad.”
For Uber, information is key, and the five-star system can provide this.
“[The system] gives you more degrees of freedom to express yourself,” Truong said. “It gives you the ability to say that it was good — but . Whenever anything goes wrong we want to know. [For example,] ‘hey that trip was okay, but maybe the music was too loud.’ ”
Making changes to reward the good
In Uber’s 180 Days of Change , the company’s campaign to enact magnanimous reforms, Truong’s team tweaked the system to encourage more feedback. If a customer gives a driver a sub-five-star rating it triggers a prompted question: Why? If the reason was beyond the driver’s control, the rating doesn’t count against the driver.
“The ratings protection feature ensures drivers are getting ratings that don’t come with guilt,” said Truong. This isn’t just a “nice” thing to do, but rather something that’s useful to the company. Uber can only leverage its mountain of review data if it’s honest. For the ratings team, Trong said, the question is, “How can we help improve the feedback aspects?”
The flip side is how to give drivers kudos, not just demerits. In the old system, a driver could only lose stars, not gain them. Truong said since a lot of the rides were good experiences for customers, Uber wanted to “start celebrating and shift to positive reinforcement.”
As a result, Uber added a tipping option to the app, finally putting the hotly-debated topic to bed. In the past, Uber had given mixed messages as to whether tipping was included in the price or not. Many customers assumed it was, with a one-size-fits all price tag — a frictionless alternative to a taxi. However, Uber recognized in 2016 that gratuity was not included in its compensation models and permitted drivers to be tipped in cash.
The built-in option, however, is a big change for drivers, who will benefit from the behavioral aspect of having a tipping screen pop up on customers’ phones. A reminder can make all the difference between tip and no tip, and the app will provide a frictionless experience for customers who don’t need to look for cash at the end of a ride.
The rating system for Uber is comparatively simple. As Truong notes, a successful, five-star ride should be the standard. Compared to a stay at a hotel, for example, a ride is a simple thing that doesn’t require a complex classification of quality. So while Airbnb, for example, may have a long road ahead in figuring out how it wants to handle a system where almost every rating is really positive — this makes sorting by ratings next to useless— Uber has a better shot at developing the perfect system.