Understanding and leveraging supply demand anomalies in the mobility industry

607 0
607 0
scooter + cars image

There are two primary variables for optimizing uptake of your mobility assets: price and placement. You can hone your strategies for both with accurate demand forecasts, yet anomalies still occur and can have major impact in this hyper-competitive industry if your fleet isn’t best positioned for them.

I’m looking forward to speaking on this topic of “Predicting and Leveraging Anomalies in Transportation Demand Management” with businesses in the Mobility as a Service (MaaS) industry at the February 2020 MOVE conference in London. While we come as experts in how events impact demand for transport companies, this roundtable will allow some of the top global MaaS organizations to explore the full range of what causes demand anomalies such as events, lack of supply, weather, discounts, and marketing campaigns (your own and competitors).

Exploring some of the key factors involved in understanding demand fluctuations in the MaaS industry will enable participants to learn from each other and tackle some of the big challenges for transport businesses.

This is a topic I am passionate about because we work with the world’s leading mobility companies including: two leading rideshare companies, one of the leading scooter companies, and one of the world’s largest automotive companies.

Most companies prioritize placement over price wars

While ridesharing companies can and do focus extensively on price and driver incentives, placement is make-or-break for every business in this space.  

Whatever you’re working with—cars, scooters, bikes, vans, boats, or something else—knowing where demand will hit ahead of time ensures you are in the best possible position to get maximum uptake and market share.

Smarter placement increases both ride numbers and profit-per-ride, and is a particularly pertinent issue for companies that are constrained by city or country regulations based on the volume of assets they can have at particular locations. If you can only have six bikes at each bike rack in an area that is about to experience peak demand, you’ll need a well-planned strategy for rapid replenishment.

It is also hard to deviate from known hotspots of demand (such as outside of a popular sports venue or conference center); however, knowing how to predict where people will be and/or where they need to get to is key to optimization.

Being proactive about placement and predicting upcoming demand anomalies means you can ensure you have assets ready, before a location shows even a hint it’s about to become a hotspot. This is critical because there is no automated way to shift your assets yet. By the time you are aware a demand spike is occuring, chances are it’s too late to capture a sizable portion of it.

We need to consider mobility’s demand elasticity when identifying anomalies

MaaS companies have the potential to address a plethora of diverse use cases. But, this also leads to a range of different competitive situations which each need to be understood to be best optimized.

For example, when a 10,000-attendee concert ends, users will prioritize immediate access to get out of the crowd before the queues and wait times kick in. A large conference will have a similar impact, as will time-sensitive trips such as getting to the airport or a key, timely event.

Other than events, severe weather can cause unexpected surges and plunges in demand – causing even frequent scooter-hire customers to pull out their phones and open up a rideshare app instead.

This is different from the more sustained, less urgent demand than when immediate discovery and use is critical for uptake.

Tapping into perfects storms of demand

It’s important to understand these differing levels of demand urgency when optimizing your strategies, especially when it comes to preparing for perfect storms of demand. These occur when smaller events cluster and combine to have a similar impact to a large, well-known event.

For example, this perfect storm in Austin, Texas is made up of events that are ranked mostly below major and significant impact (for more on how we rank events by impact, read this post). This is drawing close to 650,000 people to the area — all of whom need ways of getting around — and would easily be missed compared to a similarly sized, single event.

Perfect storms are particularly impactful for mobility companies as they will cause unexpected demand and displacement of assets.

Context: How mobility operators use PredictHQ

Transport companies use our intelligent event data in two ways:

  1. Directly in their demand forecasting models
  2. As a way to save city and operational managers’ time while ensuring they are aware of all event-driven potential demand anomalies

Most companies use it directly in their modelling. Our API is designed for ease of use with an identical schema regardless of event type and the ability to swiftly correlate historical data with verified and ranked events.

Correlation of mobility demand with events reveals three important insights:

  • What kind of events cause increase and decreases in your demand, so you can identify for future events and optimize your plans.
  • Unexpected or underappreciated hotspots that require inclusion in your strategies and potentially greater investment to capitalise on demand surges.
  • The ability to move beyond “we had a spike here last year” to pinpointing exactly why you experienced that spike and ensuring you can meet it again. Many events recur, and 85% of recurring events change location or date each time.

While many transport companies scrape Google for events or have a list of key holidays, having a much larger set of verified and ranked events means you can identify more demand anomalies reliably enough to shift assets and teams in advance to meet demand. One of my favourite moments with new customers is watching the relief and excitement from engineers and data scientists when they realize our demand intelligence data has a consistent schema and ranking methodology – saving them hundreds of hours cleaning and formatting data.

This post is focused on event-driven demand anomalies – but both this session at MOVE and PredictHQ as a company are actively exploring other kinds of demand anomaly causes. We started building our demand intelligence offering around events as one of the largest and hardest to predict sources of demand catalysts, but we’re only just getting started!

Even if you are unable to make MOVE this year, get in touch to explore how demand intelligence works for transportation networks and mobility operators. If you are coming to MOVE, be sure to join our roundtable and say hello so we can jump into what kind of events are driving your demand – and how you can make the most of them.

In this article

Join the Conversation