Questions About the GSA’s Travel Data Challenge

On February 19, 2014 the U.S. General Services Administration (GSA) announced whatThe Beat calls a “first of its kind” competition to develop travel data analysis technology. The formal announcement posted by the GSA says:

“In this GSA Travel Data Challenge, the public is asked to develop a technology-driven solution using GSA travel data that allows an agency to identify opportunities to reduce costs.  As such, GSA challenges the public to create a tool using GSA travel data that could be replicated across government to every agency, using their own travel data.”

This exercise is described as “crowdsourcing” in other publications that usually pay little attention to how the GSA administers its travel programs or analyzes travel data–one supposes because the novelty of the approach somehow makes it important.

The Travel Data Challenge raises a number of questions that are equally worthy of some attention. Here are some of the most obvious:

  1. The competition offers $35,000 for a winning submission; lesser amounts to other categories totally $90,000 in all. This is but a fraction of what such a solution would likely earn in the open market, which makes one wonder why any established developer would want to participate.
  2.  

    Participants grant the government a perpetual, royalty-free license to any and all intellectual property comprising the winning entry. A good deal for government, but a bad deal for a truly innovative developer. While the terms of the contest go on to say that “All other rights of the winning entrant will be retained by the winner of the competition,” since the rules also say that “The final tool should be in Open Source Code,” we are left to ponder how little those remaining rights might be worth.

  3. The GSA has existing contracts-holders for a variety of travel management and analysis products. Why isn’t the innovation and creativity the agency desires forthcoming from these presumably well-funded and well-compensated sources? Perhaps the agency should be questioning whether its procurement and program management practices are truly adequate to deliver the sustained innovation it seeks, or if that is not the problem, then whether the incumbent vendors are up to the job?
  4. The Organization and operation of the event give the impression that many aspiring participants are unprepared for the task. Travel data analysis and interpretation is a complex and highly specialized field. The agency has provided only the sketch of what it wants to accomplish, and many of the online questions posted on the event site indicate that an understanding of the sources, tools, and objectives of successful travel management are equally barren within the community of interest developers. A much more thorough developer briefing is needed if all sides of the contest are to avoid wasting their time.

 

Perspectives on Data Ownership: 2013

This analysis appeared March 15, 2013 in Cornerstone Information System’s “Insight & Opinion” section.


Recent popular discussions of “big data” (a surprisingly ill-defined term) are curiously silent on where these data may come from and who should decide where and how they are used. Perhaps this is because the current social media wave encourages individuals and businesses to surrender a degree of privacy (and hence control over data) in return for the promised benefits of whatever service is on offer.

While we may believe that travel data ownership questions were settled long ago, control and ownership questions are more complex than many assume and require careful review regardless of how open or restrictive data access should be.

Everyone’s In Charge

Most travel businesses you speak with will assert either that passenger travel data belong to them or that they have a right to use and distribute them essentially as they see fit.

Corporate travel managers usually maintain that, since they pay the bills, they both own and control the data. Airlines and other vendors often assume the right to use and distribute data about the use of their services, and travel management companies believe they have a degree of ownership because much of the most valuable travel data comes from their systems and exists because they expended energy to create it on behalf of “their customers.”

This travel data ownership conflict is a familiar story, but there are other less evident or considered levels:

A number of processing intermediaries including payment systems, ticket processors, GDS companies, and on-line booking tools assert a right to distribute travel data and reports for their financial benefit, apart from any direct or indirect benefit travel buyers receive. Typically this is done with individual travelers remaining anonymous, but the degree to which “anonymous” travel detail is widely available, down to specific itineraries and dates, would surprise most travel managers.

Many sources also make data available to third-party aggregators, who also operate for their own financial benefit under the assumed anonymity of individual travelers. Such companies produce an array of usage and comparative models, predictions, and similar data projects which find uses far removed from travel management.

Assumed Anonymity

I use this term to describe the broad assumption that, if my name isn’t present, whatever follows doesn’t matter. Anonymity can unravel quickly. It’s hard to argue that the kind of industry-wide data aggregations used by the DOT and others to predict economic trends are threatening, but under the care of a skilled analyst, extensive company-specific and individual travel patterns could be deduced, especially by combining multiple sources.

Interesting Questions

The extent and depth of travel data distribution and usage should at least cause travel managers some reflection, even if they decide they need not be concerned.
Here are a few specific thoughts:

  1.  complete chain of custody affecting anyone’s travel data is unknown–sometimes adequate, elsewhere non-existent. Many companies with data responsibilities have no real data security program in place that runs deeper that simply saying the right things.
  2. How is it that so many travel industry business intermediaries are selling data produced by customer activities for their own benefit? Aggregate industry analytical reports are one thing–distributing detailed raw data to third parties is another. Where did that permission come from?
  3. Have corporate travel managers looked at the type of data being distributed about their travelers in detail and rationalized it with their own company privacy and security practices?
  4. Are travel management companies comfortable with the extent of peer comparison by vendors and subsequent data aggregation that has become commonplace in the industry?

Improving Travel Industry Research

Once again a new year brings another round of what passes for industry research. Although notoriously over-surveyed, the travel industry remains awash in bad data, ill-conceived and poorly executed research projects, and self-serving studies that are relevant more to the next round of funding or the next newsletter sale than to developing a real understanding of markets and trends.

Eventually the industry may get better at labeling useless research for what it is (the trend is not positive, however), but for now the very few good studies routinely drown amidst the hyperbolae of research that can’t connect with real insight–or those that connect all too well because the result was fairly evident before the process began.

Nowhere is the problem more acute than in the online travel and social media worlds. High-priced research typically reinforces conventional wisdom and assumptions while key customer and behavior questions remain unresolved.

I’ve wondered aloud in past articles why major trade groups show such slight interest in these issues. If the online and social media worlds have such monumental consequences, what precisely could be more important to their members?

Here are a few suggestions for modeling forthcoming research projects. These are similar to suggestions I’ve made in public for 15 years, and hopefully they will help you appreciate the limitations of today’s travel research and be positioned to improve it in the future.

    1)    Broaden the Base

Successful studies need wide participation and sponsorship. Those funded and controlled by a single company or clique are not necessarily bad, but this adds complexities and concerns that are avoidable through planning and execution that strives to include more viewpoints. Addressing the needs of a broad constituency increases both value and integrity.

    2)    Sampling Is Key

Few research projects undertaken in the travel industry describe how the study sample was selected, what the resulting accuracy and margin for error are, or the size of the sample.  This is because these are among the most challenging aspects of valid research–requiring time, expertise, and money to address properly.

Most researchers simply ignore them; the resulting studies are little better than worthless.

If you peek under the covers only slightly at a surprising number of major industry studies, you’ll discover that the sample essentially self-selects. The researchers won’t explain how their conclusions in this environment are valid because they aren’t–and they can’t.

There are many parts of a study that has sufficient statistical validity to become the basis for real-world conclusions and predictions, but one is usually that a valid sample must be defined and identified in advance of the research and then the study must continue until it reaches the sample as defined.  More work than most researchers want.

    3)    Questions, Questions

Any question-based research should disclose the questions used and how these are presented.  Forming valid questions is a significant undertaking–which is frequently botched.

During 2010 I was treated to a trade conference  where the expert presented study results to show the importance of the field where he was the market leader and that was the subject of his presentation.  The self-selected sample were asked a variety of simplistic questions with many obvious answers:

“Is cost-control important in your business?”

Have you ever met anyone who would answer “no?”

When the presenter reached some study questions that were clearly silly, he remarked,

“Well, my staff assembled these questions and I should have reviewed them better.”

In other words, the presentation is a waste of everyone’s time–which also says something about the extent to which some conference organizers vet their presentations.

    4)    Seek Wide Input

Limiting control over a study to its sponsors or other “insiders” cannot but color the result as self-serving. Enlightened researchers learned along ago that the “best and brightest” often don’t work for them and they seek such talent out wherever they can find it.

    5)    Dump Hyperbolae; Focus On Quality

The industry doesn’t need another round or praise describing how great the opportunity of the day may be. What’s needed is thorough research and careful answers that relate to business concerns and allow the reader to reliably take action.

This simple definition disqualifies most of the fluff-laden e-commerce and social media studies of the past few years. There are people who know how to do real research–it’s a mystery when their input is so clearly lacking in the major reports of today.

    6)    Analyze

As there are competent researchers there are also competent interpreters who can make connections between abstract numbers and real business situations. Their work ought to by key to any research project. A study lacking informed, usable conclusions should be first into the worthless bucket.

    7)    Validate

There’s a saying that teaches thus:

“Premises that are absurd when projected into the future were absurd to begin with.”

Researchers and readers alike need to apply logical tests in order to understand the validity of a study’s conclusions.

For instance, most predictions about the fantastic growth of mobile, social media-based, or other new media travel purchases assume a level of personal computer use and literacy throughout society that is simply absurd within the time frames considered.

Clearly studies are failures when they cannot withstand the test of reasonableness.

    8)    Forget “Guru” Mentality

Research is ongoing and the “final word” on most topics will likely never be written. Successful research projects are willingly subject to critical review and are revised in light of new viewpoint and data.

A premise holding that the oracle has now spoken and nothing further may be added only highlights the underling weakness of the research in question.