NOTICE

The file is temporarily not accessible. Please try later!

NOTICE

The current file is MB in size and your download remaining (MB) is not sufficient.

Your monthly download limit(150MB) will reset 12:01 a.m. GMT on the 1st of each month.

NOTICE

Considering bandwidth cost, WELOVEAD provides limited free download service.

Please notice the following before downloading the file.

Befor Download MB
File size MB
After download MB

NOTICE

Due to busy traffic of our server for this moment, please wait seconds before you are able to download again.

Monty's Wicket Warnings

basics
Industry: Professional & Public services
Media:Cyber
Style: Minimalism
descriptioncnen
Why is this work relevant for Media?

Australian cricket fans can now watch live cricket in a totally new way.

They usually watch cricket for hours, waiting for the big plays (a typical match lasts 30 hours), then miss it as they get on with their day. They only see the action in replay. But using artificial intelligence (AI), broadcaster Foxtel predicted when the big plays would happen. Predicting live action before it happens then alerting fans across multiple media touchpoints in the moments before, intensified anticipation, ensured fans never missed the big moments, and boosted Foxtel subscriptions on a weekly basis.

Background

Australians adore cricket. It’s the soundtrack to their summer. But when Foxtel made a huge investment of AUD $600m (US $430m) in 2018 to secure cricket broadcast rights, there was a major challenge: How to get people to pay a subscription to watch cricket they had always watched for free when at the same time, under the country’s complex sport licensing laws, they could watch some of it for free on other channels?
Against this background, we had to acquire new subscribers, reduce cost per acquisition (CPA) by 40%, steal share of free-to-air viewing and become the number one channel on Foxtel.

Describe the creative idea/insights (30% of vote)

Meet ‘Monty’, the world’s first AI predictive sport commentator. He can tell you when a big moment is about to happen in cricket – a wicket falling – during a live game.

In a long, complex game like cricket, people watching live on TV often miss the big wickets. They go
to the toilet, to the beach; they miss the few seconds that really matter and have to watch the replay.

Over a typical five-day match there are a maximum of 40 wickets in 2,700 balls. Each ball lasts less than half a second. No wonder so many people miss them.

But Monty uses machine learning to predict when a wicket will fall with astonishing accuracy – 87%. As the moment approaches, he lets Foxtel viewers know, so they can tune in.

We nicknamed our model Monty after the first wicket he correctly predicted in training on historical data

Describe the strategy (20% of vote)

Monty is a machine learning model. We trained him by feeding him a year’s worth of Australian cricket team data (28,720 balls bowled by the Australian Men’s team). Instead of buying media inventory with our media budget, we bought historical and live cricket data from Opta, investing 3:1 into data against standard media.

And because cricket is so complex, each ball has 83 data variables attached to it (pitch condition, weather, speed of delivery, etc).

Once we were confident he was fully operational, we created an API output. This triggers time-sensitive adverts across video, digital outdoor and display alerting Foxtel subscribers to tune in to watch the next wicket. This maximizes their subscription. And for non-subscribers, they can sign-up to ensure they never miss the action.

This revolutionary idea used AI and data to predict big plays before they even happen. It has transformed the live sport experience in Australia.

Describe the execution (20% of vote)

Once trained, we let Monty loose on live games where he tracked 83 variables for every single ball of live play. He overlaid his training data with what happened five minutes earlier to predict what would happen in the next five minutes.

When he spotted a wicket coming, he triggered creative across pre-roll video, mobile display and outdoor billboards. The call to action urged people to tune in to Fox Cricket and watch the wicket fall. Subscribers never missed the big plays and the public witnessed how Foxtel could predict the future.

A Google integration allowed fans to ask Fox Cricket for Monty's call through the Google Assistant during live play and an API was integrated into the Fox Cricket app for instant predictions and enhanced visualisations.

For the first time ever, the viewer was not watching replays of what had just happened but was actually ahead of the game.

List the results (30% of vote)

• Acquire new subscribers: Average weekly sales increased by 18% after Monty began predicting wickets.
• Reduce CPA by 40%: CPA was 61% below average (i.e. 53% above target)
• Steal share from free-to-air viewing: Channel CH7 share of P16-54 dropped from 49.9% (CH9 was the rights holder in the previous year) to 20.8%.
• Become number-one Foxtel channel: We became number one for share, beating the full-year performance of every other channel in just three months.
• Monty correctly predicted over 1,867 wickets during the campaign with a personal best of 87% accuracy during India’s tour of Australia.
• Despite an Australia team playing far below their potential, viewership was 25% above target.
• Monty is being turned into a standalone subscription product for News Corp’s Supercoach fantasy platform.
• He is now expanding his predictive talents to Aussie Rules Football.
credit
Brand:
Agency:
Production Company:
awards
Cannes Lions International Festival of Creativity 2019
Silver Media
Media / Entertainment
Latest Updated
Loading...
Sponsor