Uber and some transportation experts once predicted that getting a ride with the tap of an app would help reduce traffic and increase riders’ use of public transportation. Instead, the opposite happened. I mentioned this in a recent newsletter. I wanted to go a little deeper today into what went wrong with the promise of on-demand rides and what we could learn from it. How can we believe that technology will help solve big problems if Uber’s great promise didn’t pan out? Here’s what more research is finding: In the past few years, on-demand ride services have been a major factor in increased traffic in U.S. cities, particularly in the downtowns of big cities. And most research is showing that the ride services have also been a significant reason for declining ridership of public transportation, especially buses. Uber and Lyft have said that people driving themselves are the biggest sources of traffic. That is true, but it doesn’t explain the surge in traffic that the services have added to cities. What went wrong? Gregory D. Erhardt, who analyzes transportation modeling systems at the University of Kentucky, told me that the companies and some transportation experts misjudged how the ride services would be used. The theory of on-demand rides was that they would be like carpooling. As people drove to work, they’d pick up an extra person or two along the way — and some money, too. But Uber and Lyft turned out to be more like taxis. Uber and Lyft, as they expanded, focused on dense urban areas, where there were plenty of potential drivers and riders. But even there, drivers spend a large percentage of their working hours roaming around without fares and clogging the streets, Dr. Erhardt said. The combination of all of these factors was more miles driven in many large and midsize cities. (Dr. Erhardt and his colleagues are soon publishing additional research into the effects of ride-hail services in about 250 U.S. metropolitan areas.) Dr. Erhardt and I talked over three lessons from this misjudgment. First, Uber and Lyft need to share their data so that cities can understand the services’ impact on the roads. Second, public officials need to steer transportation policy to encourage helpful behaviors and limit destructive ones. And third, new technology needs guardrails in place — and maybe those need to be established before its impact is obvious. The first point is that Uber and Lyft, which tend to keep certain information such as where people travel and idling times secret, need to share information with cities and researchers. “Cities are pushing hard and have a strong case that we should be able to use this data for planning and research purposes,” Dr. Erhardt said. His second point was about incentives. Some cities including New York and Chicago have added fees onto Uber and Lyft rides to make it more expensive to drive around without passengers or pick up fares in dense urban centers. That essentially nudges passengers and the companies to reduce the trips that could worsen congestion and pollution. Maybe you’re thinking, if Uber and Lyft are convenient, why stand in their way? That’s fair, but governments do use taxes and subsidies to encourage people to quit smoking or buy homes. Transportation that works for everyone doesn’t happen on its own. “Designing the right structures matters,” Dr. Erhardt said. And the third point is that policymakers may have to act early to impose new rules and requirements on new technologies. They didn’t do that when Uber and Lyft came along — because the companies fought regulation and the services were popular. But the effects of the ride services suggest that emerging transportation, including driverless cars, may need regulations early on to ensure that promises of a collective benefit don’t turn out to be a mirage.