Amazon Web Services (AWS) has developed a new suite of fan statistics to activate its sponsorship rights for this year’s Six Nations rugby championship.
The cloud computing wing of the e-commerce giant is in the second year of a multi-year sponsorship agreement with the competition, having first announced the tie-up in February 2019.
In the first year of the deal, the company’s artificial intelligence (AI) and predictive technology provided fans with seven statistical insights into a range of in-game events, including scrum analyses, play patterns, try origins, and team trends.
For this year’s championships the company is once again working with Stats Perform to collect live and post-match data. AWS will provide cloud storage and analytics before delivering fan insights to broadcast viewers in the Six Nations’ different broadcast markets.
The new statistics for this year’s championships include a kick predictor to show the probability of a kicker successfully scoring a penalty kick or try conversion; a “visits to the 22” statistic which highlights the number of occasions a team has entered the opposition’s 22-metre area; and a ruck and turnover locations statistic that will bring deeper insight to the ruck and how it is used as a strategic tool for teams.
The analysis will also include a dominant tackles metric that outlines the defensive strength and structure of the teams and “power game” metric showing the areas of the pitch where a team is dominant or lacking.
To coincide with the first round of weekend matches, SportBusiness speaks to Steven Bryen, senior technology evangelist, Amazon Web Services.
AWS has struck a number of sports sponsorship deals, including most recently with Germany’s Bundesliga. What are the company’s sponsorship objectives?
We’ve been working [in] sports for a long time, including the work that we’ve done in the US with Major League Baseball. We’ve also partnered with Formula One. Sport is one of our customer verticals that runs on AWS [technology]. Sports are also wanting to take advantage of the machine learning analytic capabilities that we’re providing to other industries. Take something like the Six Nations. This is a short championship, it lasts just six weeks, so the cloud is the perfect place for these companies to be running their workloads because they can set them up for a short period and only pay for them when they’re using them.
We’ve been working with sports across events to become the de facto data analytics provider, [to promote] the way that we can scale up per event and can provide advanced analytics at their fingertips. That encourages them to invest in the cloud and invest in other infrastructure and that’s why the Six Nations fits so well – it’s a very short tournament.
Is it a value-in-kind deal, or is there a monetary value attached to it?
It’s a sponsorship deal but we don’t disclose the financial terms of the partnership. Our objective is to work with our customers to get the best value out of our technology. We don’t do sponsorship for sponsorship’s sake; we do this where we believe we can add value to our customers and we’re a customer-focused company and we work backwards on their problems and help them solve them. We don’t just put logos on things, we only get involved in things where our technology can be a differentiator, such as bringing five new statistics to the Six Nations this year.
So the objective is to encourage more rights-holders and broadcasters to use your cloud services?
I don’t think we can share that information. What I can say is we work with many broadcasters and many reputable companies that are customers of ours and use our technology anyway. We are working with over a million active customers globally, so we’re pretty much already working with all of these large customers across media, broadcast and now sports events and federations as well.
How much of the AI and data analytics technology is provided by AWS and how much is provided by companies like Stats Perform? Is it the case that you provide the cloud platform and other firms’ technology is added on top?
Stats Perform have an in-house system hosted on Amazon cloud. The data when it gets recorded comes into the Amazon cloud, so we scale that up and down. They have their own software that runs there, but in terms of the machine learning, specifically, if you take the example of the kick predictor, specifically, that was built using our machine learning solutions laboratory team at Amazon. The machine learning solutions lab are a team of data scientists whose purpose is to accelerate machine learning and solve interesting machine learning problems. We provide resources from the machine learning solutions lab to then develop specific algorithms for the statistics that we then send to the broadcasters. In the case of the Six Nations, Amazon are providing the machine learning and its running on the Amazon cloud. So, Amazon are taking the raw data and turning it into insight.
What about the graphics and visualisations that are ultimately seen on screen and consumed by fans?
We have another partner involved called Alston Elliot who build the graphics that you see on screen. They have this broadcast software that broadcasters can use to take these insights. Alston Elliot take the custom graphics in real-time and then prepare them for broadcast.
Stats Perform stand at the beginning of the data pipeline, we add some machine learning and analytics and insight and provide the platform that powers Stats Perform’s collection of APIs and then Alston Elliot do the visualisation and overlay at the broadcast level.
What sort of latency is there between collecting the data and seeing the statistics on screen?
There can be something like an eight to 20 second delay from the event happening on the ground to the graphic appearing on screen. One of the things we’ve done with Stats Perform on the AWS platform is we’ve been able to reduce that latency by up to 50 per cent this year and will continue to improve that. Most of the remaining latency is down to satellite uplink time, so we won’t ever get exact real-time events.