2018 was supposed to be another bright year for Didi Chuxing. The ride-hailing upstart had been celebrated as a success story in China’s booming sharing economy. Six years after founding, it had become the default app for Chinese people to move around, thanks in no small part to its acquisition of Uber China in 2016. Meituan Dianping and other challengers struggled to curb its dominance and the confident, promising startup was reportedly on course for a public listing as early as 2018.
Then all the excitement ceased. Last May, a female passenger was raped and killed by her driver on Didi’s popular hiRR-Magazineh-hiking service. Just three months later, another incident shocked China after a Didi customer was killed in a similar circumstance.
The murders cast a huge shadow over the world’s largest ride-sharing company by number of rides. Celebrities rushed to social media with calls to #DeleteDidi, while Chinese authorities lashed out at Didi’s “unshirkable responsibility,” ordering it to indefinitely suspend its HiRR-Magazineh feature and ramp up safety standards.
Didi said there’s “no definite timetable” for relaunching HiRR-Magazineh, but people familiar with the matter told TechCrunch the service will likely resume later this year after Didi fully consults the public and all other stakeholders.
In a humbling move, Didi apologized and admitted that its fixation on growth had caused irreversible harm. In an open letter (link in Chinese) to the public, President Liu Qing and CEO Cheng Wei promised to “stop using scale and growth as a measurement.”
Safety, the executives pledged, would become the company’s top priority going forward. That’s despite the fact that Didi has not generated a profit since launching, but it’s apparently in no rush to shore up its balance sheet. Didi “has no timeline or schedule for an IPO,” a source with direct knowledge of the matter told TechCrunch.
Eight months in, Didi’s effort to brighten its dented image as an unsafe service has cut across its key leaders and internal departments, and it has proven a challenging journey.
“30 million trips are completed on Didi’s platform every day. That’s more than 10 billion trips a year. How do we make sure there is no incident in these 10 billion trips? This is indeed a challenge for the world,” asked Didi CTO Zhang Bo, clad in a plain blue shirt, during an interview with TechCrunch at the company’s Beijing headquarters.
While absolute safety is a nearly impossible task, there are other more tangible goals pertaining to growth and market share. As Didi brought on more shareholders in each funding round — the company has raised at least $20.6 billion to date, Crunchbase data shows — the bottom line inevitably became more pertinent.
Economic goals aside, many tech startups set out with an inherent sense of mission, or zeal, as some claim, to make the world a more efficient place. Didi’s was to shorten wait times for both cab drivers and passengers in China.
“When Didi first started, we had one simple goal. We realized taxi hailing was very inefficient. Before Didi existed, people in Beijing had to wait at least 20 minutes until a taxi pulled up. Even then, drivers often refused passengers,” Zhang recalled. “We wondered if we could overcome this sort of information asymmetry using the mobile internet.”
That was 2012. Didi has since morphed into a one-stop app for hiring cabs and private cars, in addition to a slew of mobility services like rental bikes and scooters, car insurances and next-gen technologies like autonomous driving that are still in the lab. Its global footprint now spans China, Japan, Australia, Mexico and Brazil. When the startup took over Uber’s local business in 2016, it quickly ballooned to account for 90 percent of China’s ride-sharing market.
Safety has, along the way, been somewhere at the back of Didi’s mind. Statistics from China’s Supreme Court show that crime incidents per 10,000 ride-hailing drivers are 0.048, and 77.8 percent of the time, riders are victimized. But until a fatal killing, the company’s drive to dominate the market seemingly outweighed the need for moral reckonings.
“Didi has always emphasized safety internally, but the two incidents from last year really sounded an alarm for us,” reflected Zhang.
The need to get millions of drivers “made [ride-sharing companies] lower their standards and affected safety, [including] background checks, driver training and education, experience, continuous checks, monitoring of working hours for fatigue and drowsiness, distraction by multitasking with a cell phone and passenger demands,” Alejandro Henao, a mobility researcher at the National Renewable Energy Laboratory in Colorado, told TechCrunch in an email.
“Safety needs to be a top priority for these companies but it doesn’t seem to be,” Henao added. “It’s a reflection of what I call a ‘quantity over quality’ mistake, where rushing to get volume — millions, billions of rides, passenger and drivers — compromises other factors, including safety.”
Safety at all costs
The dilemma of Didi is not unique. All ride-sharing businesses, regardless of scale, face the tough decision between responsible growth and safety.
They can of course work on both; but often they must pick one. In a world where algorithms run many aspects of our lives, every decision tech bosses make can shape our movement, consumption, mental state among many other things.
The murders revealed major deficiencies in the Didi app and its processes. One key vulnerability was the decision to outsource its passenger support system, which was criticized for failing to act on a prior complaint against one of the drivers suspected of killing. Keeping an in-house customer service team would clearly fortify the overall safety system, but it is a move that will impact the bottom line for a company that is already some way from profitability.
Didi reportedly suffered from a $1.6 billion loss last year. The firm recently shed some light on its financials during a Q&A with the public. While its core ride-hailing business charged an average of 19 percent in commissions, overall expenditures, including tax payments and driver bonuses, stood at 21 percent, meaning the segment lost 2 percent per ride.
Unlike other tech-powered offline transactions, where the encounter between the supply and demand sides is usually brief, such as food and online shopping delivery, ride-hailing passengers and drivers enter an enclosed space together for far longer periods of time.
“Before the encounter [between the driver and the passenger], neither of them can predict exactly what will happen. This kind of experience is unique [to ride-sharing], and other online-to-offline transactions almost never have to deal with the same circumstance. As such, many problems can emerge, especially given Didi’s size,” Zhao Shuai, head of artificial intelligence at Didi, told TechCrunch at the company’s headquarters.
Following the incidents, Didi announced it would spend 140 million yuan ($20.8 million) to improve its support system and set up an 8,000-strong customer service team. A flurry of other measures ensued, including the addition of a panic button linked to regional police stations. In-trip audio recording is now compulsory — the clips are deleted a week later assuming no dispute is reported, Didi assures. Drivers are asked to scan their face for identity checks throughout the day, as opposed to just once a day when they began their shift per the old setup.
The list of new measures continues, and there’s probably no correct answer when it comes to the right amount of safety features handling millions of daily rides.
“I had no confidence when I was given the job,” conceded Huang Yuanjian, who was appointed to lead a team of 30 top product managers to build out Didi’s safety feature. “No one knew how safety should be done in the mobile transportation industry, but it’s been a very fulfilling task being able to bring more safety through technology.”
When asked if Didi is concerned about the costs generated by its pivot, Zhang claimed that the company will “spare no effort to invest in safety.”
“Our investors are very supportive. They share our view that safety and long-term development go hand in hand… It’s hard to do safety well, but when the work is completed, it will inevitably become an industry standard,” the technology chief asserted.
Aside from the outsized investment in staffed customer support, Didi is also trying to minimize costs by replacing human labor with artificial intelligence technology. A team led by Zhao, who was previously one of the earliest product managers for Microsoft’s Chinese chatbot Xiaobing, is building an AI-powered assistant to aid Didi’s customer support staff. The bot is able to answer “the majority of” inquiries before pointing riders to a human agent if needed, according to Zhao. AI is also able to automatically create support tickets and log most of the conversation details for the agent.
“In the long run, passengers will choose what they believe is the safer option,” said Zhang. “We hope to not only be the world’s biggest one-stop transportation platform, but also the platform that offers the best experience and commands the most respect.”
As Didi works to appease customers who worry about riding with strangers, its new stringent rules have sparked concerns among drivers.
“Didi clearly prioritizes customers over drivers. When a passenger complains, even if it’s not the driver’s fault, Didi often assumes it is,” said a Beijing-based Didi “fleet leader” who manages more than 100 ride-hailing drivers.
Several other Didi drivers TechCrunch spoke to echoed that complaint, recounting cases of short-fused customers who take their frustration in life out on the drivers. It’s challenging to monitor what exactly happens inside a vehicle, save for determining the true nature of the dispute and resolving it. The public often overlooks the fact that driving a taxi, or joining as a ride-hailing driver, is one of the most dangerous jobs out there as it requires dealing with customers of all kinds, carrying significant amounts of cash and, at times, driving to remote areas late at night.
“The stricter the rules are, the less genuine the drivers become,” said the fleet leader. “Didi drivers have a lot to put up with.”
In response to the alleged unfair treatment, a Didi spokesperson said the platform “focuses on both drivers and passengers.”
“When passengers trust us and continue to use our services, it translates to better income for drivers,” said the spokesperson. “Our recent safety upgrades, including more stringent driver vetting and driver training programs, aim to address riders’ concerns and improve their trust in us and our drivers.”
Indeed, safety is not something that ride-hailing apps can solve on their own as it is “a three-part responsibility of regulators, companies, and drivers,” Adam Cohen, a researcher at the Transportation Sustainability Research Center at the University of California, Berkeley, told TechCrunch.
“Drivers have a responsibility to self-regulate and ensure that they are complying with all regulatory requirements and serving the public professionally,” added Cohen.
Didi is now under pressure to not only please its riders but also keep its 31 million drivers loyal and accountable. To that end, the company recently announced to set up a team of 2,000 support staff by year-end to serve, train and seek feedback from its drivers on a daily basis.
Besieged by challengers
As tightening regulations in China put a squeeze on the number of ride-hailing cars and drivers, new business models threaten to take customers away from Didi. One of them is AutoNavi, the mapping and navigation service owned by Alibaba, which is also an investor in Didi. It offers a ‘super aggregator’ app that allows customers to hire rides from an array of different apps — including Didi — in one place to, in theory, shorten wait times.
The model has already spawned followers. In late April, Tencent-backed Meituan introduced its own marketplace for ride-hailing apps after struggling to establish a major foothold with its Didi equivalent. Didi is also open to working with outside partners, according to Zhang, although he did not give further details of how such a collaboration might work.
“We don’t think Didi itself can revolutionize the massive transportation industry. We hope to be more open and create win-win situations. We are considering allowing outside automakers to tap some of Didi’s technology capabilities,” the executive said.
When asked if these new players constitute a threat, Zhang argued that Didi’s stepped-up safety measures “will be a huge competitive advantage in the long run.”
Didi also enjoys a significant network effect, which denotes that services or goods become more valuable when more people use them. A platform with more drivers, Zhang explained, means less waiting time for passengers.
Didi still holds a great lead over its rivals even after losing about 4 million monthly active users between August and December, stats from data provider Jiguang show. The app recorded 66 million MAUs in December, dwarfing runner-ups Shouqi Limousine & Chauffeur and Caocao Zhuanche, which are both backed by traditional automakers and each served around 4 million MAUs in the same month.
“Didi’s network is in itself a competitive advantage,” assured Zhang.