Jonathan Palmer, product director of core data services at King, the entertainment company most famous for creating the Candy Crush franchise, sees himself as being in a lucky position. He hasn’t had to generate curiosity about using data from all areas of the company as he said it is part of the company’s DNA.
“I’ve not had to try and create a data-driven culture. It’s always been very data-driven."
“I’ve not had to try and create a data-driven culture. I’ve helped get the most out of the appetite. It’s always been very data-driven. From the moment the business started creating mobile games there was a strong focus on having a strong vein of events about the games being sent back,” he said.
"We need as much data visibility as possible to make decisions early."
Palmer explained that the life cycle of a mobile game needs to be quick as games studios want to make sure they are backing a winner. “You don’t want to overcommit to a game that maybe is not going to hit the heights you want it to. We need as much data visibility as possible to make that decision early whether it is going to be a long-term bet or something we need to pivot in a different direction,” he said.
It is not only at inception that games data is useful to King. Once a game has been released, games performance managers continue to monitor its performance and make tweaks if necessary. The games team needs to remain aware of whether a game is too hard and they are losing too many players, or if a game is too easy and people are getting bored and leaving the game.
They say this happened with level 65 of Candy Crush Saga, a level so tricky it had user forums dedicated to it. Palmer elaborated: “It’s an incredibly hard level, it had its own cult around it. We realised it was causing a lot of people to churn. Using data, we were able to say: ‘we need to drop the difficulty of this level a bit’.”
Palmer said that after that change was made, they were able to see the effect in the data. The progression funnels display how many players play levels one, two, three etc. At level 65 there used to be a steep drop in player numbers but following the tweak, they saw more even distribution of players.
However, data analytics is not only used in games development. Palmer said that marketing, finance, CRM, and HR, are “incredibly data hungry teams.” He went on to say: “There’s no part of the business that doesn’t use data in some form to make important decisions.”
With big data analytics platform Looker, the data team has the architecture that enables different teams to drill into data that is more specific to them, as opposed to looking at “one-size-fits-all” data, as Palmer put it.
It seems that it isn’t only games that are being developed at King but data too. He said that data tooling has made enormous strides in the last few years, which has enabled them to do some “ground-breaking stuff”. They are doing some “interesting stuff” with Looker and real time aggregations for Kafka – an open-source stream processing platform - and data streaming sets, and they are also exploring more cloud-based technologies.
“King is a great culture to work in, especially from a data perspective."
“King is a great culture to work in, especially from a data perspective. There’s a high degree of autonomy. If you have great ideas and a great appetite to get it done, you will make an impact and people will back you to make an impact, which is great,” said Palmer.
Jonathan Palmer spoke to DataIQ at Looker’s Join The Tour conference.