Drivers who were previously deactivated and ask to have their cases reviewed could be reinstated based on new standards, according to internal documents and interviews with current and former employees familiar with the new system, as it reduced the responsibilities of those who disabled many of their accounts.
It’s one of several changes Lyft has made to its safety processes since going public earlier this year.
Some current and former employees say the changes pose safety risks because they introduce a black-and-white decision structure for situations that may call for human judgment. Critics assert it’s similar to cost-cutting measures taken at other companies when streamlining operations and could allow Lyft to increase its share of drivers amid a heated market share battle with rival Uber.
Lyft says the new structure is aimed at removing bias from important safety decisions and establishing a clear policy on unacceptable behavior — and not for cost cutting. Incidents can be reviewed on a case-by-case basis. And there have been no fewer deactivations of drivers on average, Lyft spokeswoman Alexandra LaManna said in a statement.
The ride-hailing giant is facing growing scrutiny over passenger safety because of allegations that it doesn’t treat women’s reports of sexual harassment appropriately. The Washington Post reported in August on the concerns of women who reported instances of sexual harassment and other types of misconduct by Lyft drivers, after which they said the company failed to take enough action. And more than a dozen women are suing Lyft in San Francisco, alleging they were raped or sexually assaulted and that Lyft let “known sexual predators” give rides and concealed sexual assault complaints in multiple locations.
The company has made a number of changes to its platform amid the scrutiny, adding new safety features like continuous background checks and a feature that automatically senses when a ride veers off course, though it said those features had been long in the works.
Lyft instituted its new safety processes in June. Safety team members now follow a set of standard guidelines to determine whether a driver has committed an offense to prompt deactivation, according to the documents and employees, who spoke on the condition of anonymity for fear of retribution.
Previously, lower-tier team members said they had the ability to identify potentially dangerous behavior from drivers when riders reported feeling unsafe and preemptively boot them for good. They had the flexibility to ban drivers for a pattern of behavior demonstrating a lack of fitness, such as repeatedly prompting customers for details about their trip or asking personal questions such as whether they live at the drop-off destination or are single.
Now the lower-level team must follow stiff guidelines that could result in drivers staying online until after a more serious incident, and recommendations to review a serious case are approved by a new team that makes the final decision based on those criteria, employees say. Those specialists, called Safety Policy and Community Compliance, make the decision based on whether deactivating a driver lines up with the standards established in the company’s new “decision matrix.” Details on those specific criteria were not available. Lyft said that patterns of behavior are addressed in the guidelines.
Lyft has not informed drivers they can ask for a review. But in documents obtained by The Post, the company tells employees that if a driver or rider had been permanently banned, “the decision can be reassessed by SPCC team members.”
Lyft already is incentivized to maintain as many drivers as possible as it competes with rival Uber, the employees said. The more drivers with access to the platform, the shorter the wait times and fares, according to its stock market filing and other public guidance from Lyft.
Lyft’s LaManna said in a statement that the company can still make decisions to deactivate drivers on a case-by-case basis, guided by the newly standardized system.
“This team follows a rigorous process to determine when drivers should be deactivated from Lyft, representing one further action toward our goal of making Lyft the safest form of transportation for everyone,” LaManna said. The process aims to cut down on the previous differences in how specialists decided on deactivations, as well as eliminating the need for critical safety decisions to rest in the hands of individuals.
Academic and legal experts say the standardization of how to remove drivers at Lyft could shield the company from allegations that its processes are unscientific or unfair. The current and former employees said there was sometimes a variation between how many drivers one specialist would deactivate on a weekly basis compared with another. The new system, which involves a computer-guided process to help determine whether a case should be considered for deactivation, aims to create consistency.
“Removing human nuance around something as sensitive around which drivers are barred or disbarred … is an effort to reduce costs and to abdicate responsibility,” said Scott Galloway, a professor of marketing at New York University’s Stern School of Business who has focused on the technology industry.
Lyft’s changes come as the company’s valuation has been sliced in half since it went public in late March and reported a $644 million quarterly loss in the second quarter. The company reported strong revenue growth in the most recent financial period, however, posting a narrower loss of about $463 million.
Stock prices rose initially in after-hours trading on the results, as well as a revised financial outlook that predicted higher-than-expected growth and revenue going forward. Lyft said its quarterly revenue grew to $956 million, a 63 percent increase compared to the same quarter last year.
In the internal documents viewed by The Post, the safety team changes were described as an extension of a model introduced earlier in the year to offer a “best experience at low cost,” a new companywide mantra.
Lyft’s LaManna said that it invests more resources and talent in safety every year. “The changes we made earlier this year were steps taken to improve the safety of the platform, and any suggestion to the contrary is uninformed and false,” she added in a statement. “They were not intended to — nor have they — cut costs for the company.”
Trust and Safety head Mary Winfield will be leaving the company, effective Nov. 8, the company confirmed. She announced the decision internally last week, saying it was a personal choice and that she did not have any immediate plans. The reason for her departure and her replacement was not immediately clear. Lyft declined to make her available for comment.
The Trust and Safety team, now based in Nashville, has been around since the app’s launch in 2012 and serves as a layer of protection for riders and drivers using the app, according to the company. It operates a round-the-clock critical response line where workers take reports from riders and drivers on safety violations, escalating many of those to specialists who conduct investigations through interviews and data gathering and undertake other efforts to determine the credibility of complaints.
Its safety regimen operates akin to a triage system, where offenses are categorized according to their seriousness: from issues such as hard braking to unwanted comments or sexual harassment to physical violence or sexual assault, according to the employees and documents. Drug and alcohol violations such as possession or an open container are deemed “zero-tolerance” offenses.
In extreme cases, Lyft Trust and Safety specialists banned drivers from the platform as a tool to maintain the safety of the app, even in situations where the credibility of offenses could not necessarily be corroborated but there were multiple allegations demonstrating a pattern of potentially dangerous behavior, the employees familiar with the process said.
That changed in June. Employees received an email informing them the new Safety Policy and Community Compliance team, would be established to review the most serious complaints about drivers and decide their fate, according to an email reviewed by The Post.
The team “was created to provide an improved and more consistent experience for cases that may require a permanent deactivation of a driver,” Lyft said in an internal slide deck.
Lyft, in its internal communications, told employees to follow a slide deck listing determining factors for when incident responses should be escalated, saying a lot of “thought and hard work” went into the process. Trust and Safety team members now follow guidelines on a computer screen and elevate potential deactivations to the SPCC team, whose members make a final decision.
LaManna said in the statement that the new team “ensures consistency and prevents bias when permanently banning drivers from the platform for serious offenses and violations of our Community Guidelines.” She added that the team follows a rigorous process.
In August, Lyft sent another email to tell its dozens of Trust and Safety employees that they have been switched from salaried positions to hourly status, according to the internal communications.
Now those former specialists, newly branded “associates,” say a layer of defense has been eroded.
LaManna said that the company has invested in the team and processes, something that has “elevated our standards and strengthened Lyft as we strive to keep drivers and riders safe.”
The changes “may have unforeseen consequences,” said a former executive, who spoke on the condition of anonymity because of the sensitivity of the matter. “At the end of the day, you’re making a judgment call, so you need to make sure that you have people who are able to understand the process look at the data.”
Still, the former executive said, Lyft’s reasoning for instituting the program could be sound.
“You’ve set up a program where you’re basically automating some of the manual things that we did,” the person said. “Maybe you’re weighing judgments, you’re putting in a more scientific approach. That would allow you to lower costs. You would assume the same fidelity of the program, that’s very feasible.”
NYU’s Galloway said Lyft would ultimately be forced to reconcile with a system that deals with matters in which there are “a million shades of gray.”