Barriers to Innovation in Travel Forecasting

It could be argued that every technical field has its barriers to innovation, but barriers are particularly acute in travel forecasting. If photography innovations were adopted at the same rate as ours, many of us would still be using Instamatics. A good friend of mine recently described the regional model in his locality “the best black and white television money could buy.”

Please bear with me for a personal story. By 1986, as an associate professor and 12 years beyond my PhD, I reached a sobering realization that not a single thing that I had done had been implemented into practice.         At that time, I egotistically believed that at least some of my research was practice-ready, but nobody was listening. The solution for me was highly elaborate – an unlikely option for other researchers, since travel forecasting software has now reached such a high degree of complexity that it would be surprising to find more than a handful of people who could successfully adopt the dual roles of researcher and software developer.

So if you were a person with an innovation today, there are huge barriers to be faced, listed here in no particular order.

Publication. You would probably want to get the idea published somewhere, but unless you can convince 3 or more referees that the research is significant, you may be stuck with the options of a conference presentation or maybe self-publication on the Web. The refereeing process is hazardous for practical research, as we have discussed previously. The chance that somebody visiting your poster or sitting in your podium session might actually want to implement your innovation is miniscule.

Local Agency Resistance. Public agencies are notoriously resistant to change, but there are notable counter examples. I have often been amazed at what certain people have accomplished within their own agencies, but I suspect the most successful of them got what they wanted through good fortune, force of personality or, more likely, both. If you are working in a public agency, being persuasive and headstrong are assets. However, if you are an outsider, there is minimal chance that you could sway the agency to your way of thinking. Many public agencies have an irrational (IMHO) desire to defend the integrity of the last planning study, and thus to advocate the correctness of the current travel model. Attempts at public shaming don’t seem to work.

Agency Data are Tied to an Existing Model. Agencies have huge investments in data. The value of an innovation must be demonstrated beyond conclusively when a massive new data collection (or reformatting) effort is required.

Software Platform. Software licenses can be expensive, but importantly, staff are reluctant to abandon software and algorithms they are comfortable using. Nobody wants to incur data conversion costs and headaches. Some software products have steep learning curves, so switching spends much human capital.

Standardization. Standardization smothers innovation. Good intentions by state officials to standardize models can lock everyone into mediocrity.

Software Developer Resistance. Most people doing travel forecasting are dependent on software developed by somebody else. If you want to innovate, there is a very good chance you would need to convince a software developer to write new code for you. New code is expensive and it is logical that a developer will weigh the costs of the new code against future product sales. Developers won’t touch an innovation unless it is in the public domain. Sometimes an innovation is incompatible with the architecture of the software or violates some philosophical principle, so implementation cannot happen, regardless of good intentions. Absent of the ability to write new code, your innovations are limited to parameter tweaking.

Funding Agency Resistance.  You might need funding to demonstrate the relevance of an idea. FHWA and FTA are the two federal agencies most involved in travel forecasting and they employ many smart and dedicated people who see their role as fostering innovation. However, I have gotten the impression that their research is limited and targeted. Getting their attention can be difficult. RITA amply funds the UTC’s, but access to a UTC is restricted, and research needs to adhere to a UTC’s theme. Inspirations don’t happen because funding is available. The idea needs to come first, then perhaps funding will follow. NCHRP, et al. occasionally fund projects related to travel forecasting, but the dynamics of review panels often work against true innovation.

Lack of Places to Advertise Innovations. Discussion forums and the TFResource are not really suitable for advanced concepts that have not received extensive vetting.

Peer Reviews are Inherently Limiting. Peer reviews are great for bringing models up to contemporary standards, but reviews are restricted to what the peers know. Speaking from my own experience, peers are reluctant to advocate for an idea unless they are pretty sure it has already worked well somewhere else.

Clearly we need ways of breaking down barriers; the easy ways to get ideas into practice are not available to everyone. However, let me at least propose a few feeble strategies that are less ambitious than writing your own travel forecasting package.

Implementation as Research. Some of the most rewarding research is showing how an existing idea can be implemented for travel forecasting. Examples abound. Publishing a pure implementation paper can be tricky, but often doable with the proper level of salesmanship.

Talk to Software Developers. Based on my own desires, I can speculate that most software developers are very interested in any free ideas they can implement relatively quickly. Don’t expect software developers to pore through hundreds of articles just to find your idea. Some ideas that seem trivially small can be quite worthy.

Talk to the Feds. Remembering that federal employees are busy people with their own tastes and inclinations, they are usually willing to talk about an innovation, unless you are begging funding for it. Start the conversation first, then see what happens. Federal employees are often in a position to carry an idea forward, when they find it sufficiently interesting.

Talk to Agencies. I am referring to agency-to-agency communication. Agencies like to hear success stories from other agencies. Be proactive at model user-groups. Submit papers to conferences likely to be attended by other agencies.

Self-Publish. We need to be careful with self-publication. It is easy to embarrass oneself. And we don’t want to violate copyrights. However, some items deserve to be placed on the Web. To do this, you will need access to a web page. Things to self-publish include: technical reports done for an agency; thesis or dissertation; and elaborations of shorter articles that might have gotten into conference proceedings.

I would be interested in hearing about any other thoughts you might have.

Alan Horowitz, Whitefish Bay, December 2, 2015