Senior Manager for Rider Services and Policy
Micromoblity is still in its infancy and the processes and programs we use to manage riders and vehicle assets are evolving. My colleague, Bob Mallon, recently wrote about the dire consequences for micro-mobility companies if they cannot build trust in the community: trust that assets are available, that they work correctly, that they are where they are supposed to be. Part of creating trust is providing our riders with all the tools they need to succeed. When this happens, riders arrive safely at their destinations and assets are left secure and available for the next rider. At Zagster, we are using and developing cutting-edge tools to identify both good and not-so-good rider behavior. We use our deep knowledge of the assets we operate and the on the ground realities experienced by our riders to connect core rider values (I like riding this scooter and want to ride again) with rider actions (I will make sure the scooter is returned to the designated scooter parking area for the next rider).
Ridesharing companies experienced a similar period of rapid expansion and new product introductions along with all the associated risks. The successful ones employ rider behavior analysis as one of many tools to mitigate risk. Of course, if you hop in an Uber, the app is collecting all sorts of data from the times you typically call a car to the most popular destinations in a neighborhood. That data allows companies to understand how riders are using their products and how rider needs and company assets and efficiency can be balanced. The flip side of this is rider scoring, where the behavior of a rider in the app and during rides can be used to provide specific education and even enforcement. That data makes up your rider rating.
At Zagster, this is a concept we improve upon daily as we bring world-class micro-mobility fleet management software to the communities we serve. We’ve built software and created processes that identify specific rider behaviors and combinations of behaviors linked to bad outcomes. It is important to take a step back here and understand that 99.9% of riders aren’t looking to steal scooters or other assets. Rather, multiple factors from a bad cell signal to an ancient phone to just good, old-fashioned impatience are out there in the real world to cause problems. That’s where we come in.
My background is in New York City government, data-driven policy, and community organizing. When I took this job, I saw an opportunity for us to better understand our riders. One of the first things I did was rack-up more than 500 rides on our Pace bike share bikes in under four months. This field work left me with a comprehensive understanding of the asset itself and how riders use the asset. Then, I developed a rudimentary set of aggravating and mitigating factors that seemed to be the strongest indicators of ride outcomes (an asset being left unsecured, etc.). More on this later, but that became the framework for the concept of a Zagster Rider Score.
Rider Score is exactly what it sounds like, a numeric score assigned to an individual rider that attempts to map that rider’s ride behavior through the lens of factors that generally result in assets going missing. Again, this score is wholly behavior based and generated from actual ride data. The current version of this scoring system bears little resemblance to the simple system I initially designed. Using machine learning, we were able to develop a model to identify a constellation of over 100 predictive factors that linked riders identified as struggling, confused, or in rare cases, malicious and connected to missing assets. Again, there is no silver bullet here, no one single factor that causes poor micro-mobility asset outcomes. Different Rider Score factors and factor combinations interact to produce different scores. At the most basic level, a “bad” score is grounds for a rider suspension, much the same as a high-speed trip through a school zone should land a car driver in serious trouble. Of course this is an important service. But, the real value of a system like this rests in its ability to identify small and specific rider behavior combinations that have the potential to combine, snowball, and worsen over time into something more serious. These can be addressed and corrected before this happens with a simple reminder--an email, additional app screen, or educational video. After all, remember your own first bike ride. We can all benefit from a little helpful instruction once in a while.
The benefits of our Rider Score are both enormous and applicable across all micro-mobility assets. Just like my 500 Pace bike rides set me down the path to predictive bike share rider analytics, the same principles can be easily applied to an e-bike, e-scooter, or the next new vehicle yet to appear. The concept is the same. At Zagster, our comprehensive knowledge of product, rider, and the operational challenges present in the real world allow us to tailor our approach to the unique challenges of different micro-mobility assets, markets, and conditions. This approach guarantees that at the end of the day, these bikes, scooters, and other vehicles are available to the riders and communities that rely on them. It is fleet management at its best.