Seeking Data in an Imperfect World

<B> Seeking Data in an Imperfect World</B>

By Christopher M. Brittin

<hr><I>Christopher M. Brittin is president of International Software Products in McLean, Va.</i><hr>

In the coming perfect world, all data required for corporate travel managers to make informed purchasing and policy decisions will be instantly available, fully detailed, absolutely correct and totally consistent. Even accepting the rate of change known as Internet Time, rumor has it that this goal might not be reached next week. Nevertheless, the decisions will need to be made next week. The art of business often requires choices and judgments based on incomplete information.

There has long been a disconnect between the data available to travel managers and the information they require. The industry has been slow to respond, partially due to a focus on information presentation (reporting systems), rather than collecting, distributing and consolidating the underlying data. Existing reporting systems often permit data to be imported in any format, as long as those formats match the import specification of the reporting system (an approach similar to the color choices for the Model-T). Corporate travel departments purchasing these systems often become involved in the business of data transformation and consolidation. Furthermore, already-overburdened industry MIS departments have little incentive to prepare yet another report or extraction, or simply lack the technical resources to respond.

To address these and other issues, the industry appears to be taking a long overdue look at the problem of information interchange and data standards. Recent articles in BTN (<I>BTN,</I> May 17, June 7), have highlighted efforts currently underway to formulate a standard, including the OpenTravel Alliance XML proposal and the Prism Group's Xport specification. Both make a valuable contribution to the data standards discussion: the OpenTravel by embracing an emerging international standard, and the Prism Group by spotlighting data quality. But how should we evaluate the merits of these and other proposals?

Although not exhaustive, I believe that the following criteria are central:

Is the proposed standard truly open? Can anyone contribute to the discussion? Is it controlled by a single industry segment or even a single company?

Can the standard be implemented easily and at low cost? Can business take a phased approach to deployment to reap benefits soon?

Is the standard flexible and extendable? Do we really think we now know all the data elements that will ever be required or all the processing requirements? If additions are required later, can existing systems continue without modification?

Does the standard embrace transactional and analytical requirements or is it better suited to one model?

Does the standard clearly separate logical content from physical format? Are any limiting assumptions about interchange methods made?

Is the model truly global, or does it suffer from a national or regional bias? Are even simple items, such as varying international date representations, ignored? Americans, in particular, seem to be remarkably provincial in our outlook. Can the standard adapt to the culture or will the culture be forced to adapt to the standard?

Can the standard deal easily with incomplete data or missing data elements?

It is worthy to aim for rich data sources, but at present, not all sources can provide equal detail. Missing data does not necessarily invalidate the elements that are present.

In my experience, travel managers are also deeply interested in preserving choice. A standards development effort that focuses only on the largest players may have the undesirable side effect of reducing choice. Successful open standards have in a large part survived because they are ultimately enabling in the truest sense, and they tend to increase opportunities for competition while reducing costs.

Finally, participants in efforts to develop and agree on a standard should not overlook the decision maker's need for global information now. The industry does not now, and will not in the immediate future, have a single data interchange standard. While moving toward the standard, flexibly dealing with data in disparate formats from multiple sources remains a necessary bridge, and ultimately a vehicle for rapidly moving to any standard.

Certainly the data consumer must be well informed about the content of the data, but the requirement to deliver that data in a timely, cost-effective manner is paramount.