How Prediction Markets helped Channel 4 predict TV success more accurately

In today’s hyper-competitive broadcasting landscape, where choice paralysis abounds and budgetary strain places even more scrutiny on commissioners’ hit rates, predicting a new programme’s chance of success has become more challenging than ever.

For Channel 4, traditional commissioning methods that rely more on gut instinct, experience, or more conventional survey data, are no longer cutting the mustard. This is where our Prediction Market tool – leveraging a projective approach to early stage concept testing and optimisation – comes in.  An extensive trial this year has highlighted how Prediction Markets can be a game-changer for commissioning teams, and help broadcasters like Channel 4 make smarter, data-driven investment decisions.

The Challenge: Navigating a Crowded Media Landscape

Fragmented viewing habits, younger audiences migrating to digital platforms, and intense competition from streaming services have drastically reshaped the broadcasting industry. As Channel 4’s Insight Business Partner, David Benabo, explained in front of a packed audience at the MRG conference:

“With audiences having so much choice, they’ve become more risk-averse. A new programme must be close to perfect to be considered. People look for any reason not to watch something.”

Commissioning a new TV show involves significant risk. Development and production costs can run into the millions, with no guarantee of success. Even well-made programmes can struggle if they don’t immediately captivate their audience and stand out amidst the plethora of compelling options available to viewers.

The challenge they posed us was unequivocal, whilst simultaneously exciting and slightly terrifying: was our Prediction Market tool more accurate in forecasting a show’s success than other approaches, and could it give them greater confidence in their commissioning decisions?

What Is a Prediction Market?

Prediction Markets are a gamified approach to using online surveys to forecast outcomes at scale, leveraging the “wisdom of the crowds.” Drawing on decades of academic research, including studies from Harvard, Stanford, and the Nobel Prize-winning work of Daniel Kahneman, its earliest applications can be found in financial and political predictions, and it has since expanded into TV and entertainment.

Here’s how it works:

1. Wisdom of the Crowds: Instead of asking participants about their personal opinions (‘how do you rate this new show concept?’) or claimed viewing intentions (‘how likely are you to watch?’), we ask them to predict how others will react to a new show concept.
2. Gamified Betting: Respondents place virtual tokens or cash on their predictions, and are able to ‘spread’ their bets, based on their gut reaction
3. Adjusting for confidence levels: The use of gaming dollars as a proxy for confidence levels allows us to weight  predictions, ultimately yielding more accurate forecasts than traditional methods

Crucially, this approach shifts the focus from personal preferences to projective judgment, mitigating the behavioural biases attached to self-reporting and providing more considered reflections on what might resonate at a broader, more mainstream level, and why. Here’s an example of someone justifying how they predicted a new reality show idea might do, if it were aired:

Methodology: Testing the Test

Channel 4 have been using this tool with us for years, but it hadn’t necessarily been widely adopted within the business and needed a strong proof point to convince all commissioners; it was time to put it under the spotlight. They provided summaries of a range of shows already commissioned that were due for a 2024 release – allowing for almost immediate comparison against viewing figures. These were run through the tool, and predictions were compared to actual viewing data post-broadcast. Success in market was defined as a show achieving above-average ratings for its slot.

Key metrics from our data set used for the trial included:

• Top Box Predictions: Shows predicted to perform “Well above average” on a 5-point scale by a certain threshold were deemed likely to succeed.
• Self-Reported Viewing Likelihood: For comparison, we also used traditional survey approaches asking respondents if they would personally watch the show, again on a 5-point scale

9 out of 10 concepts predicted to do well went on to perform well on air

The results were remarkable:

• Prediction Markets achieved 93% accuracy in identifying whether a show would succeed, far outperforming the traditional survey methods (only offering 50% accuracy).
• Asking respondents to predict its performance was significantly more reliable than asking about their personal preferences, proving that “the wisdom of the crowds” could be a powerful tool for broadcasters.

David emphasized the importance of this milestone:

“We essentially wanted to test the test. This was a high-stakes trial, and the results exceeded our expectations. Prediction Markets give us confidence to make decisions grounded in data, not just intuition.”

The Broader Impact for commissioners: harnessing data for a better hit rate

This trial confirmed what researchers at Iowa University and beyond have argued for decades: Prediction Markets are not just a theoretical tool but a practical, real-world solution. Their applications go beyond entertainment, extending into areas like product development, packaging, and marketing strategy.

For broadcasters, the implications are profound. Prediction Markets offer a way to mitigate risk, accelerate the failure of weak ideas, and focus resources on developing hits. As David succinctly put it:

“A graveyard of well-made TV exists because shows didn’t quite do enough to be added to viewers’ schedules. Prediction Markets help us avoid that fate.”

If you’re ready to take the guesswork out of concept testing, improve your accuracy for predicted success and unlock the potential in the wisdom of the crowds, drop us a line.