Product Owner of Business Intelligence Lukasz Korbolewski shares how King is using Anodot’s AI Analytics to proactively address business incidents in real time, to minimize revenue loss and identify opportunities.
Presented by Lukasz Korbolewski
Lukasz Korbolewski is product owner for business intelligence at King. He is a BI expert who loves elegant data models and modern analytics solutions. King is the leading interactive entertainment company for the mobile world. King’s products have captivated gamers worldwide, most notably Candy Crush, Farm Heroes and Pet Rescue.
Video Transcription
King is a mobile gaming company. And our most successful and most recognizable franchise is Candy Crush. At the moment, we have 18 games in production, Candy being, obviously, the most popular one. We also have Bubble Witch franchise, Farm, Pet Rescue, and some other minor games. Currently, I’m working in the incident management team and I’m a product owner of business intelligence for the team.
What it means is that imagine you have a team of data analysts or people who basically need to use data to investigate things. So, the incident management is investigating incidents, and looking at the loss that we have when incidents are happening. It’s really important to find out what’s going on, what incidents do we have, and what’s the root cause of them.
First of all, this is the detection tool. What we’re looking here is to minimize the loss, the impact on revenue. So, we need to spot incidents as soon as possible. With Anodot we are looking at minimizing the average time to detect, at the same time, have accurate alerting on anomalies that are happening. Using the Anodot we are seeing time to detect of 20% of the previous. So, it’s 80% improvement to the previous year.
Initially, we started with basic metrics like game starts and game ends and the transactions. But then we expanded this to more interesting metrics. So, at the moment we are monitoring 18 KPIs for each game, and this is quite a lot of metrics that gives us a pretty full picture of what’s going on. But we also split this into different dimensions. So, we look at which platform the game is performing on, also which build that is, and in which country. Eventually, it gives us around 20,000 times services that we’re monitoring.
Looking at business KPIs is much more challenging than just looking at IT infrastructure. The noise or the anomalies are much more subtle. So, the tools that we were using so far weren’t good enough for that. We had different areas such as maintenance, or how much data we can put in. One of the most important ones was detection, detecting different scenarios that we’d find. Anodot was the best in picking these up. Apart from that, we also looked at how does it deal with false positives, because we learned that this is one of the most important things not to be alerted too often. And Anodot was providing the whole interface where a business logic for differentiating between incident and anomaly could have been implemented by using just simple settings.
We decreased the time to detect, the most important thing for us. Secondly, we do have real-time anomaly detection in place, so we feel safer. We know that we’re gonna get an alert if anything is happening. The most significant change, for me, is that we have something in place that we can depend on. And this gives us, you know, a comfort and confidence that we are able to acknowledge these events. We can respond to them in a quick manner. And as a result, we can just perform better.