Delphi Forecasting | The Panel of Experts | The Process | Sample letter | Summary | Links

Organisations in fast moving technical markets must continually evaluate technological advances. There is an ongoing risk of a substitute technical solution which may obsolete current offerings also the risk of backing the wrong horse while emerging technologies are still incubating. It is a typical management situation where decisions must be made before the facts can possibly be known.

Lack of action is unacceptable even when backing the wrong horse can damage your prospects. The old adage that there are only three types of company, those that "make things happen", those that watch, and those wondering "what happened!" was never more relevant.

We would probably all like to be one of the first group of companies "making it happen" failing that, and I am sure most organisations will fail to be in that category, the Delphi forecasting technique can assist you at least to become an active participant in the second group of companies that are "watching what is happening" and based on active watching you can hopefully take the required decisions before it is too late.

Membership of the third group "those wondering what happened!" is not an option unless you like to bury your head in the sand and live "one step from chapter 11".

Delphi Forecasting

The Delphi forecasting process is a systematic, iterative predictive research method based on independent inputs from a panel of experts. It measures the degree of consensus among the panel regarding future events.

It is a good technique that I believe includes some little tricks to make it more accurate than you might at first expect.

The Panel of Experts

The panel should consist of people in key engineering / management positions who are in a position to make informed judgements on an industry`s future and themselves influence future trends.

For a particular industry, a good panel would probably include at least one key individual in every major player. The research organisers and administrators will have to decide both the number of panellists they can handle and the number needed to ensure progress can be maintained if, for example, some panellists are unable to respond to a questionnaire in the time allocated.

The key factor in the individuals chosen for the panel is not just that they are experts in their field, but that being key players in the industry concerned they are actually in a position to influence the future outcomes you are seeking to predict with this research.

The Process

The research method involves repeated rounds of questionnaires where responses are re-circulated in the form of median and interquartile range measurements so individuals can reconsider their opinions based on the responses of the panel as a whole.

The initial questionnaire must include items of interest to your firm and areas you feel will be of interest to Panel members. Members of the Panel will be able to propose questions in the course of the research.

Another key point making this technique interesting is that panel members remain anonymous from each other at all times during the process to ensure that the results are not influenced by group or peer pressures. In face to face discussions or focus groups a bully or dominant personality may hold a much greater sway than their expertise or influence should allow, this technique avoids that risk completely.

The research can be administered by post or email or perhaps even via a website or down the telephone, the important consideration being that panel members are kept remote from each other during the process and individual comments and responses are not identified to any panel member.

Tasks to establish an ongoing research system based on the Delphi method of forecasting.

1. Draw up a list of proposed panel members, approach them informally to gauge their reaction to the proposition. Panel members should be industry experts, working in the industry, able to predict the future and to exert their influence to help their predictions come true. There could be problems of direction if both technical and commercial / marketing people are in the panel.

2. Draw up the first questionnaire to deal with issues of interest to your organisation that the Delphi panel are qualified to comment on.

3. Approach the prospective panel members with the proposition and perhaps the first questionnaire. The first questionnaire must have been exhaustively tested by colleagues, to ensure;

It is clear and easy to follow and return, perhaps with already addressed return envelopes or return email.

It will require approximately 10-15 minutes for filling in. (more might be too intrusive for the benefits on offer to the respondents)

The completed questionnaires must be easy to analyse, and to aggregate results for re-circulation (e.g. responses are quantifiable etc.) Useful measures for describing the aggregate results are the Median, Interquartile range and Averages. Questionnaires can be identified internally or coded so research administrators know who completed each one.

1.The first questionnaire .. will present the first group of questions and prompt panellists for their initial responses, comments, and, proposals for further questions or topics to be covered. Space must be set aside for proposed further questions or future topics. The first new questions would be fed into the second questionnaire at which time the first questions will be being looked at for a second time.

2.The second questionnaire .. will present the first group of questions again, with the same opportunity to respond but will also show the aggregate results from the first time each question was asked and any unidentified open ended comments next to the repeated questions. In this second questionnaire, the respondents are asked to answer the first questions for a second time, in the light of the other panellists grouped responses. This second questionnaire will also hold a group of questions that are being circulated for the first time created by panellists suggestions on the first questionnaire. Questions could be grouped in batches according to whether they are on their first or second exposure.

3. The third communication might include a report of the results from the very first set of questions which have now been looked at twice and will now have probably been exhausted. In any case, additional questions on similar subjects will have been raised and will be part of the current questionnaire no 3. This third communication would include the third questionnaire containing two groups of questions, those on their second iteration and those on their first. The questionnaire would still be designed to produce proposals for new questions and topics to keep a fresh supply of relevant up to date issues for the panel to address.

4.Continuing, ongoing repeated research whose method of question renewal should ensure relevant and up to date issues are explored.

Notes .. The timing of the programme cannot be too demanding on panellists or administrators, Probably a questionnaire a month would be a fair balance, assuming they are concise, relevant and interesting to the panellists. Research can continue as long as benefits are being gained and panel members are willing to continue.

Sample letter of introduction:

Dear ... ,

Following our meeting / telephone conversation, I am writing to further explain a research initiative that I think will interest you.

Because of your position and experience in the XYZ industry, I hope you will be intrigued to know the consensus view of your industry peers regarding the future use of ABC. What I have to offer you is the chance to test your theories on the future with other industry experts from which a picture of the industry into the next century will we hope emerge. I am sure this picture is of as vital interest to you, your organisation, and the other experts I am contacting as it is to our organisation which wishes to play a central role in this future.

I would therefore like to invite you to be a member of the panel of experts in this Delphi study. Being a member of the panel will involve no more than completing a 10 - 15 minute questionnaire every month or two. You will have the opportunity to propose your own topics and questions for the panel to consider. If you agree to be a member of the panel of experts, the identity of the other panellists will be kept secret from you. The intention of this is simply to remove the influence of peer pressure and group dynamics from the research which will be conducted postally.

The results of the study are to be used as follows.
* [insert your reasoning here or modify mine below]
* To generate future scenarios for the use of ABC in the XYZ industry.
* To be issued to members of the Delphi panel.
* For comparison with a report on the XYZ Industry.
* To be used to consider the viability of a European study.

If you agree to help us in our research please indicate this on this letter and return it in the envelope provided, (or please complete the enclosed questionnaire). If however you would like any further information please do not hesitate to contact me.

Summary - "A final thought":

I hope you agree that is in interesting technique, the question of panel motivation is an important one and you may find individuals are concerned with competitive interests.

You may be tempted to use the auspices of your industry association to add credence to your research, in fact as long as your organisation has an expert on the panel if you can persuade your industry association to administer the research you get the results without the pain!

How much hassle have I just saved you with that!

[Author mark@sticky-marketing.net 04 April 2001]

Top | Home | Articles | Glossary | Sitemap | About
Copyright 2001-2004 sticky-marketing.net