Sportsbet turns to data insights to meet live betting demands

Sportsbet is using a mix of historical and recent data to forecast staffing levels needed to handle customer service and live betting volumes in 15-minute increments.

The online bookmaker has spent the past three years optimising workforce planning and management for its contact centre operations.

It operates three contact centres in Melbourne, Darwin and Malta.

“Our core customer service centre is operated out of Melbourne,” head of workforce planning James Moore told Customer Contact Week A/NZ.

“In addition, the legislation that we're governed by at Sportsbet dictates that we're required to operate our phone betting centre out of the Northern Territory … so our Darwin centre takes all of our phone betting, but they also take a bit of customer service overflow.

“Then, to leverage 24x7 overnight support for customer service, [we have a] Malta team [that] can [also] flex up the amount of resources that we've got when there's events that are on that dictate high call volumes.”

Sportsbet receives between 25,000 and 30,000 customer service contacts a week, “with a heavy split towards live chat, which is a preference we want to continue to push” as an online bookmaker, Moore said.

The live phone betting contact centre, meanwhile, sees “between 70,000 to 120,000 calls in a standard week when events are on”, though that dropped to between 30,000 and 40,000 a week at the height of COVID-19 sports cancellations.

Moore said that 33 percent of live betting volumes came in between midnight and 7am, “which is quite high for an overnight period.”

Volumes were heavily dependent on the attractiveness of live odds on offer.

“The phone betting centre is a really unusual contact centre to be planning contact for because forecasting is extremely difficult,” Moore said.

“For example, if a favourite is winning a match really easily and the odds aren't really great, we probably won’t get too many calls around it.

“But if there's a match where a real underdog has taken a bit of a lead, and the odds are great and people think the favourite might come back, we'll notice a big spike in contact.”

“We're heavily driven not by just historical data but we have to predict what type of match it could be, because if we get that wrong and it's a popular match and we don't have enough people on, the customers who want to bet will have to wait a long time in a queue.”

The company had put a 90 percent service level target on live betting contacts.

“We've got high grade service targets,” Moore said. 

“We want most of our customers basically getting answers straight away. 

“When live odds change, they change quickly, so if a customer's waiting too long in a queue and they don't get their odds, then it's a really poor customer experience.

“We try to minimise that as much as possible.”

Getting staff levels right

Moore said he had been brought into Sportsbet three years ago to “evolve” the bookmaker’s workforce planning strategy.

This covered not only operational tactics used within the centres but also creating a better understanding of the “commercial benefit” of getting forecasting and service levels right.

“In the strategic space, I wanted to take a more rounded view of understanding the importance of how workforce planning can be an important impact to how the business succeeds,” Moore said.

Moore said Sportsbet saw data and insights as critical to workforce planning.

Data, he said, could be used to identify opportunities to improve certain “customer journeys”, which could lead to a fall in contacts and ultimately an increase in revenue.

“If we're getting 25,000 to 30,000 contacts a week, and we can get it down to 20,000, that's a real measure of success because our customers aren't contacting customer service and they're on our app betting [instead] because that's what they want to be doing,” he said.

More recently, the focus has been around using technology - Sportsbet did not specify which - to automate more of the handling of certain types of inbound contact, and find trends in the reasons why customers are making contact.

“When you're looking at contact, really understanding ... why contact is happening and what you can be doing to support your customers [is important],” Moore said.

“What we're doing well at the moment is we're trying to automate the information that customers see, so if they're contacting us, the information is getting told to them before they need to speak to an agent about it. 

“We’re also automating their self-service journey on the app for little things like if they need a password reset. Having our app as advanced as possible to be able to self-service those things is pivotal. 

“Additionally, having strong data mining to be able to see things are going well, or if they're going poorly, so you can put strategies in place to reduce that impact is also really important.”

Moore said that Sportsbet is also starting to deflect customers out of live chat by automatically providing answers to common questions.

“What we try to do is deflect as much live chat contact as possible, so we put an information banner at the top of the most common questions and if that answers their question that will immediately end the chat,” Moore said.

The company is also using a mix of historical and recent contact, event and betting information to forecast the staffing levels for its centres at any given time.

“How do we forecast for staff given every day is different?” he said.

“We've got a sound amount of long term historical information that we use. We put more emphasis on the recent information per event, so we will literally build our forecast around the events that are on and we'll get an extrapolated volume of the amount of bets and customers that will be online during those events.

“We've got our forecasting down to 15 minute intervals - so we staff to 15 minute intervals. Going any lower than that is probably a bit too [granular] for resource requirements.”