The analysis behind the decision to form INFORMS

Editor’s note: As part of our coverage of the 25th anniversary of INFORMS following the merger of its predecessors ORSA and TIMS, we invited Robin Keller and Craig Kirkwood to contribute a “Last Word” column based on the cost-benefit analysis they conducted leading up to the 1995 merger. 

We were chief financial officers for The Institute of Management Sciences (TIMS) and the Operations Research Society of America (ORSA), respectively, during portions of the period leading up to the merger to form INFORMS, and we were actively involved in analyses to support decision-making about the potential merger. Keller and Kirkwood [1] provides a definitive record of those analyses, and we will provide a few additional thoughts looking back from 2020.

First, the multiobjective decision analysis approach described in the above-referenced article provided a neutral framework for clarifying, analyzing and discussing issues involved in redefining the relationship between TIMS and ORSA. With this approach, objectives are specified, and potential alternatives are identified to address these objectives. In a full quantitative multiobjective decision analysis, a metric is specified to measure the degree of achievement of each objective, and a value function is determined to address tradeoffs among the objectives. There were many differing points of view among the involved leadership over the several years that planning occurred, and this approach was generally accepted as a useful way to address the issues.

Over the period from 1989 until the merger in 1995, the specifics of the decision analysis evolved, but the approach continued to provide a valuable framework for addressing the issues in a way that ensured all points of view were included in the analysis. This is in contrast to planning approaches seen in some organizational planning situations where a subcommittee develops a single alternative and presents it for possible modifications by the larger group. That type of approach was less appropriate here because of the long planning period, over which there were changes to elected leaders of both organizations. In this changing environment, multiobjective decision analysis provided a stable platform for new people to get up to speed on the ongoing planning process, and within which they could readily see that their specific concerns were being addressed.

Second, the financial part of the analysis provided a key insight: While most leaders of ORSA or TIMS were also members of the other organization, this was not true for the general membership of each organization. Because joint members paid the same dues to each organization as people who were members of only that organization, joint members were providing a financial subsidy to single-society members. This was because joint members received few additional services for their second membership dues. Therefore, unless expenses for the merged organization were significantly reduced, the dues for a substantial majority of members would have to go up in a merged organization. There was concern that this might lead to a drop in total membership. That did happen to a certain extent immediately after the merger, and it took several years after the merger to work through the financial implications of the organizational change.

Finally, the analyses clarified that many of the issues facing the field in the period leading up to the merger were not tied to the existence of two separate organizations, and therefore would not be immediately impacted by the merger. However, it is hard to see how we would have been able to as effectively take advantage of new opportunities, such as the emergence of analytics as a major management interest, with the cumbersome governance structure that existed with two separate organizations.

We took a neutral stance toward the various options while we had important roles in analyzing these during the period leading up to the merger, but we both became convinced by the analyses that a merger was the best option. We have been very pleased to see the progress that INFORMS has made over the past 25 years, and we look forward to even greater advances in the future.

Postscript: ORSA had two classes of members, usually referred to as full members and associate members, and only the full members could vote. Many ORSA members did not bother to go through the process to become a full member since it did not give them any additional services. As a result, the full members were a small portion of the total membership, and informal polling indicated these tended to have a stronger attachment to the concept of ORSA as a separate organization than associate members. Therefore, some ORSA leaders were concerned that the voting ORSA members might never approve a closer affiliation with TIMS. This was addressed by making a relatively non-controversial effort, led by Robert Armacost, to convince qualified associate members to apply for full membership. Then, as we recall, a change was approved by the expanded full membership to grant voting rights to the associate members. After this, the final vote to merge with TIMS to form INFORMS was passed overwhelmingly by ORSA members.

Reference

1. L.R. Keller, C.W. Kirkwood, 1999, “The Founding of INFORMS: A Decision Analysis Perspective,” Operations Research, Vol. 47, No. 1, pp. 16-28. Full access at https://pubsonline.informs.org/doi/10.1287/opre.47.1.16.