Defining the Peskin Ratio, and why (some) scooter networks fail

Teaching cities and scooter operators to work together.

Nobody can agree what success looks like for scooters.

Some – myself included – dream of a future where streets are populated with a healthy mix of efficient, clean, and equitable public transportation and a slew of last-mile micro-mobility services, whether that’s bikes, scooters, or maybe even these funny things called “PodRides.” Why not.

For most cities, that’ll never happen in my lifetime.

It makes me sad, but it’s true. Take San Francisco: between the 490,000 registered cars, 70,000 trucks and 20,000 motorcycles, 275,000 dedicated parking spots which span the length of California’s 840-mile coastline, and both legislators and residents that seem hell-bent on hating anything that isn’t powered by gasoline, a micro-mobility utopia is far out of reach.

Some cities are trying, but their efforts are understandably limited. Santa Monica built miles of dedicated bike lanes, installed scooter parking zones, and even unanimously passed the first flexible cap on scooter deployments, letting Bird, Lime, Uber, and Lyft deploy more vehicles as demand grows.

But, you know. They keep breaking. They end up blocking sidewalks for disabled citizens. People fall off them. Oh, and they’re covered in poop.

How can we define short-term success, given a micro-mobility-first future is far out-of-reach?

To frame what I hope we can all define as success for new methods of transportation, we need to define a new metric: the Peskin ratio.

The Peskin Ratio is the ratio of “failed rides” to “successful rides” an average user experiences attempting to use a transportation service.

It’s named after Aaron Peskin, SF’s most prominent anti-tech member of our Board of Supervisors. He led the charge against scooters, passed an Uber and Lyft tax, and bizarrely is trying to ban in-office cafeterias at tech companies in an attempt to get tech workers to “go outside” — or, you know, order delivery.

A Peskin Ratio is relatively easy to calculate.

Let’s say your typical commute is in a car that you own, service regularly, and park inside your locked garage. For every thousand trips that you attempt, your car might only fail once (eg. a drained battery).

Your Car’s Peskin Ratio = ~1 failure / 1000 trips = ~0.001

The more your Peskin Ratio, the less happy you’ll be with any particular type of transportation (and more willing you’ll be to change).

I used a log scale on the x-axis. This is typically far more compressed to the left.

Let’s put it another way: if your car broke down every ten times you needed to drive to work, you’d find a new way to get to work.

Uber and Lyft have a similarly low Peskin Ratios. If you live in a major city, thanks to surge pricing and thousands of drivers, a private car is available most anytime for some nominal price. For every thousand trips that you attempt, the app might only crash or fail one or two times out of a thousand – typically rare events when the entire service goes down or a mass migration of people temporarily eats up the entire supply side.

Public transportation in most cities have low Peskin Ratios. Sure, you might not like the smell, the train might be ten minutes late, but it’ll do the job.

Even personally-owned bicycles or scooters have low Peskin Ratios in most cities. Pending theft, most bikes and scooters work every day in most weather.

What does this have to do with scooters?

The Peskin Ratio of small scale vehicle deployments is embarrassingly high, frustrating users to the point where people simply give up on scooters.

(An aside: I’ve almost exclusively experienced scooters in San Francisco, so most of my examples skew to the Bay Area. These issues are widespread committed by most operators, not just those in the Bay.)

Scooters aren’t consistently placed in the morning. Regular scooter users should expect to find scooters in similar areas every day. Bird, Lime, Spin, Skip, and Scoot — all had too many “scooter drop zones” across the city, meaning only a small and ever-changing number of drop zones are ever filled.

There aren’t enough scooters to go around. In cities where local governments have passed fleet caps, it’s nearly impossible to provide consistent service to the entire city. Back during the heyday of Bird and Lime‘s deployment in SF, each had thousands of scooters and still couldn’t service the entire city.

Now, Skip & Scoot are held to a maximum of 1,250 scooters combined. Neighborhoods who were promised coverage (and the associated foot traffic) are already complaining. This is squarely the SFMTA’s fault.

Most scooter company’s apps are woefully bad, driving up Peskin Ratios.

  • In the early days, most companies didn’t let users quickly report broken or missing scooters. Scooters have always broken, been vandalized, and “disappeared,” but not providing a clear way for users to report leads to a vicious cycle. Dozens of users will continue to look for the same scooter, get pissed, and churn.
A Peskin Ratio of 1. 😮
Peskin’s living his best life. 🤦‍♂️

The stakes here can be relatively low — a failed ride isn’t necessarily the world thing in the world, but their failure is literally injuring riders. Broken scooters aren’t taken offline, so riders will continue to use them.

(From personal experience)
  • Scoot doesn’t use QR codes on their scooters, confusing users who accidentally unlock one scooter… and get charged for hours. Then they threaten users with hundreds of dollars in fines for their bugs.
  • Back when they first launched, Skip didn’t disclose when they take their entire fleets offline (They’ve since fixed this issue). Excited users would open the app, just to discover an empty map.
So we should pick different scooter companies, right?

Even the best scooter app wouldn’t solve the problem. It’s not enough.

No scooter company can succeed when cities and companies optimize for overall fleet usage rather than specific target Peskin ratios.

Right now, the main metric cities use to judge a scooter network’s health is number of trips per scooter per day. By forcing scooter operators to optimize for usage, you drive vehicles to high-traffic areas where an overwhelming number of users compete for a minuscule fleet. Most users – those who miss out on a scooter– will leave the experience disappointed and churn.

High Peskin ratios (large numbers of failures, few successes) piss people off.

Instead, force companies to optimize for low Peskin Ratios for specific, limited populations in specific neighborhoods:

  • This minimizes the number of bad experiences with scooters. By disqualifying riders that we know will be disappointed with poor availability up front, companies correctly set expectations for everyone.
  • This aligns the incentives of companies (profits) with cities (sustainable shift from cars to scooters, helping the environment and protecting citizens from dangerous car accidents).
  • This minimizes the number of poor scooter riders. A more-engaged network of riders can be held to a higher standard than a less-engaged network. Getting banned from Amazon is a lot scarier than getting banned from that store you visited once a few years ago.

Instead of optimizing for hyper-growth, micro-mobility providers should focus on engaging a core set of daily users to drive delight, not toss their vehicles in busy areas in the hopes they’ll squeeze a couple extra rides in.

This is much easier said than done. Optimizing solely for low Peskin Ratios might drive more sustainable growth, but would discourage scooter operators from experimenting with new neighborhoods or fleet designs. A blended model is likely best, with some leeway to allow limited tests.

What does this look like in the real world?

  • Tests to qualify for scooter access. Input your daily commute & how frequently you’d like to drive a scooter to see if you qualify.
  • Automatically reserved scooters in the morning. Providers will increasingly reserve individual scooters to guarantee availability. We’ve already seen Bird move in this direction with morning delivery.
  • Less strewn scooters in touristy, busy areas. With more targeted networks, scooters will live where people work, not where people want to go walking and exploring on the weekends.

I don’t doubt that this would make scooter trips more expensive. Less rides means higher fees. But when smaller vehicles can help the environment, free up space in cities, save lives, and even shift our city’s culture for the better, we should be happy to pay the price.