On paper, San Diego seems like perfect e-scooter country. It’s got terrific year-round weather and no shortage of young residents. But in January, the micromobility company Lime became the third micromobility company in just five months to yank its shared scooters off the city’s streets. Like Jump and Skip—the other two companies to leave—Lime claimed that onerous regulations drove its decision. (Lyft, Bird, and Spin remain.)

Multiple San Diego city council members expressed their disappointment, with one blaming the city itself. “This is a classic example of what happens when government overreaches,” councilman Chris Cate tweeted. “[Lime] will not be the last company to leave.”

Indeed, the wave of belt-tightening sweeping the scooter industry suggests more retreats could follow—not just in San Diego, but nationwide. E-scooter companies are consolidating and pulling out of big cities like San Antonio and small ones like Charlottesville. More exits are likely in the coming months, potentially leaving cities in the position of places like Bakersfield. Since Bird left last November, that California city of 380,000 residents has no shared e-scooter service at all.

That would be an unfortunate outcome to those who (like me) think e-scooters can offer a clean and space-efficient way to move around. The future of these often-polarizing devices is particularly tenuous in less-dense cities and suburban areas, where elected officials may not realize that market conditions there are much tougher for micromobility operators. The wrong approach to regulation can become an e-scooter-killer.

But leaders of suburbs and less-dense cities can increase the likelihood that their e-scooters survive. The trick is to align regulations with market realities, because what works in Chicago (11,960 people/square mile) probably won’t fly in Chattanooga (1,223 people/square mile).

To understand why, let’s review some of the ways that density affects scooter demand and operations. Consider two hypothetical cities: Cramville and Scattertown, which have an equivalent number of e-scooters permitted, similar infrastructure, and equal populations. The big difference is that Scattertown is triple Cramville’s land area, giving it a third the population density.

Each square mile of Scattertown will contain fewer people than Cramville, which means that residents of Scattertown will, on average, have to walk farther to reach the nearest e-scooter (some won’t bother, opting for another mode instead). The average journey distance within Scattertown will also be longer, which means a lower proportion of trips will fall within the three-mile range where e-scooters are most popular.

Each of these factors suggests that Cramville’s e-scooter fleet will be more heavily used than Scattertown’s. As noted by Kyle Rowe, Spin’s head of government relations, a successful e-scooter service needs “a lot of demand for people to move one to three miles, and in less-dense environments that demand is simply lower.”

To go a step further, imagine that both Scattertown and Cramville require scooter renters to leave their vehicle in one of the dedicated parking zones spread throughout each city (assume they have the same number of zones). In Scattertown, this requirement will deter more potential riders, because their destination is more likely to be too far from the nearest parking zone. Demand for e-scooters in Scattertown relative to Cramville just fell even further.

To switch gears, now consider e-scooter operations in the two cities. Scattertown has a bigger geographic footprint than Cramville, which requires companies to spend more time and money redistributing and charging their vehicles. “The time necessary to pick up 10 e-scooters will be longer in a less-dense part of the city, so it’s more resource-intensive,” Spin’s Rowe says. In this case “resource-intensive” means a higher labor cost, and more time, money, and gas consumed driving around to collect and redistribute the fleet.

But even if e-scooter demand in Scattertown is lower—and operations more expensive—than in Cramville, a different set of regulations and permit fees could level the playing field. Local officials in less-dense cities can take a number of actions to help e-scooter businesses succeed.

First and foremost, cities can limit the number of companies issued an operating permit. Spin, for example, has chosen not to even enter less-dense cities like Indianapolis and Dallas that place no limitations on the number of e-scooter operators. Similarly, “Cleveland doesn’t have the same demand as San Francisco,” says Rowe. “You’re better off having fewer vendors in a place like Cleveland, which can both boost utilization of devices and reduce operational costs of providers.”

Katie Stevens, a senior director of government relations at Lime, feels similarly. “In a suburban environment, we’d probably need to be the exclusive provider,” she says.

Officials in less-dense places can also help by giving e-scooter users freedom to park near their destination. Mandatory parking zones can work well in dense Asbury Park, New Jersey, where most e-scooter riders will have only a short walk. But in a more suburban environment, long walks to or from a parking zone are more likely, and a bigger problem.

Making parking zones optional rather than mandatory is one way to go. In markets like Charlotte and the Cleveland suburbs, Spin says it gives a $0.50 discount to riders who leave their vehicle in a dedicated zone, but riders are free to leave their vehicle elsewhere on a sidewalk if they prefer. Lime’s Stevens says that Phoenix’s requirement that riders leave e-scooters only in parking zones contributed to her company exiting that market in January, though she says Lime might return if the city changes its mind. “Our hope is that they would permit an incentive system [in place of] geofenced requirements, while allowing more of a free-floating model in less dense areas throughout the city.”

A major business challenge for shared e-scooters in suburban environments is the lower usage of devices. The typical suburban e-scooter will generate less revenue for its company—but it will also hold its charge longer. Local rules requiring scooters to be removed from the street every night to be charged and “rebalanced” can be a bad fit in suburbs; the vehicles may have plenty of battery power remaining, and e-scooters off the street can’t be used by customers. (Stevens cited Phoenix’s requirement that e-scooters to be picked up between midnight and 5 a.m. as a factor in Lime’s decision to leave).

But the bluntest instrument available to alter the economics of e-scooter deployment is the fees cities charge operators. When the city council in Raleigh, North Carolina, raised the proposed fee from $150 per scooter to $300, Bird and Lime both chose to leave the city.

It’s possible that leadership in more-sprawling places may decide that these kinds of compromises aren’t worth it—that the only scooter deployment they are willing to tolerate is one that involves mandatory parking zones and daily device rebalancing, for instance. If so, officials shouldn’t be surprised to see scooter companies walk away.

In that case, there is one last-ditch option for a community that finds itself scooter-less but really wants to scoot: subsidize the operators. A harbinger of things to come may be Santa Cruz, California, which ran a pilot last year subsidizing JUMP e-bike trips downtown. It’s not a big mental leap to imagine public subsidies expanding from e-bikes to e-scooters—both offer alternatives to polluting, congesting automobile trips. Most public bikeshare systems wouldn’t exist without public support.

That scenario might sound improbable, given how many people still see e-scooters as sidewalk-cluttering nuisances, not valuable mobility services. But as the industry matures, infrastructure improves, and the market for greener transportation options expands, more cities may be willing to help e-scooter operators find a path to viability. If so, city leaders would be wise to understand the bottom-line effects of their local regulations.