Sporting Group is both a B2B and B2C company that facilitates betting and bookmaking in 23 sports. Football makes up half of its business but it also prices for lesser-known sports like handball which is popular in Denmark.
It is known in the B2B context as Sporting Solutions and has been around since 1992 years, providing fixed-odd prices for around 50 partner bookmakers globally. Consumers, however, would know the company as Sporting Index which has been enabling spread betting for the last 10 years.
The prices are generated by taking data from a third party like betting solutions company Perform and running it through mathematical models that have been developed and tweaked over the years, allowing customers to bet on a wide variety of outcomes.
“The mathematical models generate the probabilities of various outcomes across the different markets, whether it is total goals markets, player goal minutes markets, corners, yellow cards and red cards,” said chief technology officer Peter Wallis, taking the example of football.
Then Sporting Solutions translates those probabilities into odds, and its partners then put a margin on those odds and push them out to their own customers. These odds need to be generated and moved quickly; Wallis explained that data from Perform will come from a scout sat in the stadium who inputs events like corners and red cards as they happen. End-to end from the scout to the B2B customers, it takes four or five seconds.
The problem with managing data in this fast paced business is that it is very “peaky,” according to Wallis. With football comprising half of the business so a lot of activity takes place on Saturdays around 3pm GMT.
“Everything suddenly ramps up to be 200 to 300 times normal load. Every time something happens, like a goal, that will trigger a whole set of new prices across all those 380 markets. Multiply that by the number of fixtures across all the different leagues, all happening concurrently, and you can see that data does explode,” he said.
This results in the generation of around two billion prices a year. To deal with these immense surges in activity, Sporting Solutions has made its pricing infrastructure as elastic and as scalable as possible. Nine years ago when it was first built on-prem, it was containerised and virtualised so the ability to scale was limited. It now has a cloud-first strategy with its fixed-odds pricing already migrated to the cloud but its core pricing engine moving up there in the future.
The biggest complaint Wallis is likely to face from customers is about slow prices coming in two to three seconds later than they should. This could be caused by a massive DDOS attack or an outage at Amazon or Cloudflare. When this happens, Wallis and his team will look at the various dashboards and restart various components or try to give more resource to whatever is struggling. To avoid problems becoming full-blown crises, they have alerts set up, including dynamic thresholds which will let him know if any activity is outside the parameters of normality.
With regard to one of the platforms he uses, he said: “Sumo Logic gives us the ability to pull all the servers and log files together in one place and analyse that across 750-odd servers so we can follow a transaction through from end to end.”
Prior to January 2019, Wallis and his team were using Splunk but migrated away to have better out-of-the-box functionality, allowing them to process 200 times more data than before.
Approximately £8.8 billion was spent on betting and online gambling in the UK in the year to September 2018, which proportionally is 61% of the £14.5 billion spent in the gambling industry.
With mobile phones making it eve easier and quicker to place a cheeky wager on your favourite team or player, in a sports-mad nation which is the birthplace of 10 well-known games including cricket and curling, the two billion prices handled by Sporting Group is only going to go up.