The Delphi technique

What is it?
The Delphi technique has been described as ‘a method for structuring a 
group communication process so that the process is effective in allowing a 
group of individuals, as a whole, to deal with a complex problem’ 

Where does it come from?
It has its origins in the Cold War in the 1950s when the Rand Corporation, 
funded by the US Air Force, was trying to find a way to establish reliable 
consensus of opinion among a group of experts about how Soviet military 
planners might target the US industrial system in an attack and how many 
atomic bombs would be needed to have a specified level of impact on US 
military capability. This was the original ‘Project Delphi’.

What is it used for?
Fifty years later, it is widely used for more peaceful purposes, but with the
same underlying rationale: to establish as objectively as possible a
consensus on a complex problem, in circumstances where accurate
information does not exist or is impossible to obtain economically, or inputs to
conventional decision making for example by a committee meeting face to 
face are so subjective that they risk drowning out individuals’ critical 
judgements.

It is a family of techniques, rather than a single clearly­understood procedure, 
but the typical features of a Delphi procedure are an expert panel; a series of 
rounds in which information is collected from panellists, analysed and fed 
back to them as the basis for subsequent rounds; an opportunity for
individuals to revise their judgments on the basis of this feedback; and some 
degree of anonymity for their individual contributions.

The ELTons is certainly a complex problem requiring structured decisionmaking!
The format we have used for the last seven years has a panel of six
or seven judges, working at a distance, with all communication by email
through a moderator. There are two separate stages, shortlisting and judging,
each consisting of an initial round that elicits panellists’ comments on the
entries or the products, followed by one, two or three rounds in which
panellists nominate their preferred entries. The number of rounds depends on
how quickly a consensus emerges.
Panellists send their responses to the moderator, who collates them and
circulates them anonymously after each round, as the basis for the next
round. The panellists have at each stage a full record of what comments and
nominations other panellists have made, but they do not know who made
which comment or voted for which entry. Nor do they know the final result;
like the rest of the audience at the awards party, the judges themselves do
not know the outcome until the envelope is opened!