How to Keep Judging Fair? Cannes Has an Algorithm for That Source: Dave Plunkert
Every year the stakes get higher at the Cannes Lions International Festival of Creativity, and the pressure on judges intensifies as the competition between agencies, holding companies and countries to win Lions increases.
There will be 40,133 entries this year viewed by 366 judges, who will cast 300,000 votes in total. So how can a festival ensure that this vast awards machine operates as fairly as possible?
Philip Thomas, CEO of Lions Festivals, has an algorithm for that.
Three years ago, the festival introduced a system to monitor the voting patterns of every individual jury member, analyzing how he or she is voting compared to the rest of the jury.
More Coverage >>
Rightly or wrongly, allegations of block voting have always dogged Cannes juries. "In the past it was around country pride, so there was a lot of talk many years ago around cartels and blocks of countries voting for their own work," Mr. Thomas said. "Now, as the Network of the Year and the Holding Company of the Year awards become very important, the focus has shifted to voting in these areas. This is the question we're trying to address with this algorithm."
The most recent serious scandal was in 2012, when WPP and Omnicom Group traded accusations of block voting by those holding companies' agencies on the media jury, where each group had six judges. Without naming names, WPP CEO Martin Sorrell was quoted in the press saying some judges were apparently pressured to vote for certain entries. An Omnicom agency exec retaliated by saying that WPP agencies were "briefed to kill" Omnicom agencies' entries. After the festival, senior media agency leaders conferred with the organizers, and the media jury was restructured the following year.
Using the algorithm across all the juries, the data can be configured to show how a judge's voting patterns relate to their own network, to other networks, to holding companies and to particular countries. Presented as a traffic light system -- with red indicating cause for concern -- it can also be calibrated to show by what percentage a judge has been voting outside of the norm.
Every evening, after judging is finished for the day -- or in the morning if a session has gone on far into the night -- Mr. Thomas looks over the data in search of anomalies that might indicate consistently biased voting.
"The odd anomaly here or there isn't a problem," Mr. Thomas said. "Because broadly speaking one individual juror can't influence the outcome. What we're looking for is themes and genuinely consistent voting that is not in line with the others."
Philip Thomas
So far there haven't been any incidents of consistent block or individual voting either against other networks, countries or holding companies -- or in favor of them.
"It gives us a great deal of comfort. The integrity of the voting is almost certainly the single most important thing for this entire business," said Mr. Thomas. "We provide an environment that is as neutral as possible -- we choose our juries very, very carefully in terms of the balance of holding companies, networks and agencies, and none of the entries has the name of the entrant company on them."
Judges, Mr. Thomas said, also take comfort in the algorithm: "It helps people who might feel a pressure to try and help their network win, or it can help people to see familiar work with a fresh pair of eyes. So it frees them to judge totally independently and to be what we want them to be, which is the absolute gatekeepers for the very best work."
A second measure to encourage impartiality, introduced around the same time as the algorithm, was to do away with the show of hands in the final votes to award trophies. This round is now done on iPads.
"The majority [of jurors] much prefer to do it privately, because you don't get that peer pressure. Imagine sitting in a room and the majority of people put their hand up -- the human instinct is to put your hand up, too," he said.
| }
|