Recently, I was asked to present on new DRT performance metrics, one of two presentations I gave at TRB’s International Conference on Demand Responsive and Innovative Transportation Services in Baltimore. (These presentations can be found on our What’s New Tab). As they say, timing is everything, and in this case, the FTA had just issued proposed revisions to NTD DRT reporting the week before, asking for comments. Here’s the link to the Federal Notice where the proposed revisions are stated and from there you can provide a comment. https://www.regulations.gov/document?D=FTA_FRDOC_0001-0840. I took this opportunity to heart, and submitted the following comments. I urge you to comment as well, because in my opinion, the proposed revisions and the general requirements of NTD data reporting for DRT, as discussed below and in my presentation, still fall woefully short.
Comments on Proposed NTD Revisions – Submitted to the FTA by Will Rodman on May 15, 2019
1. Comments on the Proposed Revisions
a. New Type of Service for Reportable Taxi Service
The proposed revisions as stated do clarify one type of transit-sponsored DRT service:
- The proposed revision, which would be newly labeled as “Demand Response – Taxi” describes DRT service where taxis are integrated into the service mix and dispatched centrally, that is, they become a non-dedicated resource and, in some cases, dedicated resource for scheduling and dispatching from the agency, primary operations contractor, broker, or call and control center contractor.
However the description fails to include the following:
- In Los Angeles over 50% of the service is operated by “ADA paratransit certified” taxi drivers. Note that many of these taxis provide service as dedicated runs, with trips scheduled to the taxis the day before. Other taxis provide service “on the fly” with trips dispatched to them on the day of service. In Boston and Dallas, most of the taxis currently used are used in a dedicated fashion, where the software schedules to these runs the night before. Thus, the proposed revision only covers the fact that taxis were used and that trips were dispatched centrally; it does not include cases where taxi runs are used in a dedicated fashion with trips scheduled the day before. This is an important distinction, as it better describes the service mix between dedicated and non-dedicated DRT service, and not just what kind of vehicle was used.
- The proposed revision also does not include DRT service where taxis are integrated into the service mix but dispatched by the taxi company or companies. There are ADA paratransit and transit agency led coordinated paratransit systems where lists of “taxi trips” are electronically sent to taxi subcontractors for dispatching. There is also at least one ADA paratransit/senior DRT system (in Ann Arbor) where same day trips are booked and dispatched by the taxi contractor. Neither of these service models are specifically mentioned in the proposed revision.
- Such “taxi-based” integrated services are also being provided by other types on non-dedicated service providers besides taxis. For example, in New York City, 30% of the ADA ADA paratransit trips are being provided by black car/livery operators, presumably operated by “ADA-paratransit certified” drivers (to borrow the LA term). It is also entirely possible that in the future, TNC drivers will be recruited and “ADA-paratransit certified” by third-party entities (as is being planned in Nashville) to in order to expand the resources available for non-dedicated service. Under the proposed labeling, where would these be reported? The labeling as “Demand-Response – Taxi” does not seem to apply to either instance.
- By omission, the proposed revisions suddenly eliminates transit-sponsored taxi-based subsidy programs from reporting. Many transit agencies, for example, offer a taxi-based subsidy program to their ADA paratransit customers as an “alternative” service, and have been doing so for years. (See the definition of alternative services in TCRP Synthesis 135, which was approved by the FTA legal counsel serving on the project panel.) These services, while available to ADA paratransit customers and which do fall under the general provisions of the ADA (for equivalent service, for example), do not fall under the guidelines of ADA paratransit service. They are not integrated with the ADA paratransit services; they are on entirely different platforms whether or not the requests come into a centralized call center (such as in Denver) or directly to participating taxi companies. In either case though, the trips to taxis are dispatched by the taxi companies and not be the transit agency, operations contractor, broker, or central call and control contractor. Supplemental taxi-based alternative services, as offered to ADA paratransit customers, are vastly different from integrated taxi services, and the cost per trip is vastly different as well. AS I point out below, such alternative services should more appropriately be lumped in with other alternative services, such as those provided with TNCs. Indeed, several alternative service programs use both taxis and TNCs. Here it is not important for reporting whether an alternative service trip is provided by taxi or TNC; it is important that it be reported as an alternative service.
b. New Types of Service for Demand Responsive Service Provided by Transportation Network Companies
The proposed revisions do a good job describing one but only one form of “microtransit”, now recognized in the industry as on-demand DRT service for the general public, but falls short in that it specifically limits this reporting to TNCs. Microtransit is now being provided by the agencies themselves, by operations contractor (many using their paratransit customers, but some using different contractors, TNCs and taxis.
One of the failings of the proposed revisions is it does not distinguish what the TNCs are being used for: microtransit (for the general public) or alternatives services (for ADA paratransit customers – see above), or possibly in the future, ADA paratransit as mentioned above. Unit costs could be vastly different, and the data, if lumped together, would not be very illuminating.
A second failing is that it may not capture microtransit that is provided by other means, which can vary greatly. For example, the cost of providing microtransit with a transit agency’s paratransit contractor is likely to be vastly different from a microtransit using TNCs.
2. Related Comments on NTD DRT Data Shortcomings
The general failing of NTD data is that it focuses on the type of vehicle that is used to provide the service and not the type of service it is providing, and as such, the DRT data is not very useful for peer comparison analyses, which besides as an input to funding calculations, is how NTD data is mostly used. Consider the following examples:
- A DRT service that provides ADA paratransit service
- A DRT service that provides ADA paratransit and senior and/or human service agency trips
These will likely have very different productivity and unit costs, but the DRT data would not illuminate this key difference. And the industry is plagued with comparisons being made between such systems, relying only on NTD data.
Here’s another example:
- An ADA paratransit service using dedicated vehicles only
- An ADA paratransit service that also used non-dedicated service providers (RVHs measured differently)
In these examples above, the one system using non-dedicated services, if orchestrated strategically, is likely to have a very different and higher productivity and lower unit cost than the service with 100% dedicated service. But for the reasons mentioned earlier, this might not be obvious from the NTD data.
One of the reasons for the difference is because non-dedicated service provider (NDSP) labor tends to be less expensive (e.g., no benefits), while agency/contractor driver labor costs with benefits have been shown to reflect up to 70% of the operating cost structure of dedicated paratransit services. But whether or not the system uses NDSPs at all is not particularly evident from the NTD data.
Moreover, in cases where they do use NDSPs, the productivity is going to be inflated anyway because the FTA has said its okay to report NDSP revenue hours as just the live passenger time. So, the fewer the hours, the higher the productivity. Thus, comparing the productivity of a 100% dedicated service with a “blended” system using NDSPs is an apples to orange comparison.
Here are four more examples:
ADA paratransit service with a service mix of 85%/15% (i.e., where NDSPs are
used to serve 15% of the trips) vs. one with a 50%/50% service mix.
- An ADA paratransit service with an average trip length of 8 miles vs. 12 miles?
- An ADA paratransit service in an area with light traffic congestion vs. very heavy traffic congestion?
- An ADA paratransit service that has 15% wheelchair trips vs. 30% wheelchair trips?
The first bullet above provides yet another apples-to-orange comparison because, here too, the productivity and unit costs will be very different but this information is, again, not in the NTD data.
The second bullet above compares two systems with very different average trip lengths, which will likely have very different productivity and unit cost. Now, in this case, NTD does provide some mileage clues but you have to know that the NTD ridership figures include PCAs and companions, and because of this, you really don’t have an accurate stat for average miles per trip.
The third bullet above compares two system serving areas with very different traffic congestion, but the NTD does not account for this. (There is an easy way to do this, as offered later in this comment.)
The last bullet above compares two systems with different demands from customers who use wheelchairs. Because of the extra time it takes to load and unload a customer in a wheelchair, significant differences in this type of demand can have a profound different in productivity and unit cost. But this not evident from the NTD data either.
And one last example:
- A DRT service that provides an ADA paratransit service with no other real options
- A transit agency which also offers an alternative service (taxi-or TNC-based subsidy program) for ADA paratransit customers
As mentioned in the initial set of comments to the proposed revisions, the unit cost of an alternative service is vastly different from the cost of an ADA paratransit service, but if they are not distinguished in a meaningful way, the NTD data can be very confusing, and ultimately unusable.
A step in the right direction might be to distinguish between ADA paratransit trips and other trips that might include premium service trips, alternative service trips, coordinated trips sponsored by 3rd parties, and general public trips on microtransit services. However, I think this is still broad for both categories, as it does not distinguish between dedicated and non-dedicated service, and the “other” category contains services that are very different.
What makes much more sense to me is to breakdown trips and other data by type of services and then by trip type. For example:
fleet: ADA paratransit trips vs. non-ADA paratransit trips
- Non-dedicated fleet: ADA paratransit trips vs. non-ADA paratransit trips
- Transit-sponsored alternatives services (for ADA paratransit customers)
- Micro-transit (general public, on-demand) services
Many may think this would be an imposition, but virtually all software systems in use today can generate this split out of stats and costs for each of these four categories and subcategories.
For Productivity, trips per revenue vehicle hour can be misleading for the reasons mentioned earlier. Instead of or maybe in addition to trips per revenue vehicle hour, I offer the following two new productivity measures:
- Direct miles per RVH, which for peer comparisons or comparisons of service providers in different zones or regions, helps correct for differences in average trip length
- Direct travel time per RVH, which for the same purpose, helps correct for differences in local traffic congestion
And for these two metrics, paratransit software systems that use Google Maps or an equivalent in the dispatch functions can get 100% reporting for both of these, and in cases where you don’t have that kind of a system, you can go into Google at the time of the trip with a statistically relevant sample and do your own. Note too that these two new productivity metrics are now being used more and more in Europe.
And lastly, I also suggest reporting the breakdown of trips requiring a wheelchair accessible vehicle (WAV) and trips that don’t, another data point that virtually all software systems are able to generate.