iGaming 2023: The Key to Frictionless Data Management

With the sheer volume of customer data available to iGaming operators, many are considered to be in a privileged position. The real challenge comes from the effective management of this data, as it’s key to maximising the customer experience at scale and facilitating sustainable growth in the business.

With new, innovative technology now available in the market, the options for increased efficiency are boundless, yet, some industry commentators believe we are not taking full advantage of the opportunities on offer to take our sector to the next level of engagement and subsequent growth.

We caught up with Claudia Heiling, COO of leading data science agency, Golden Whale, to hear her thoughts on the current standards of data management in iGaming and, more importantly, how they can be improved going forward.

With data collection and automation being a key area of tech development for the online casino industry, how can businesses like Golden Whale help operators streamline their manual processes?

“The answer to this is two-fold. The first part is the data handling and engineering. This often involves collating information from different systems – including legacy systems – so you first need to build a connection layer to draw it all into a common pool. This pool needs to be designed with machine learning in mind, so it has to be presented in a way that allows neural networks to build meaningful data sets.

“The second part involves model orchestration. Everyone says they have the best models in terms of making predictions and segmentations, but most of these are black box models and the operator or game producer can’t delve into their inner workings. At Golden Whale, we have a dedicated layer for model orchestration and competition, so you can see all the outcomes from similar data sets. This makes the user more independent from single sources and helps them build an evolving system around their modelling approaches. As we have a lot of domain know-how from our backgrounds in gaming/data science, we’re able to select models that can predict the target value our clients are looking for.

“With our system Foundation, this can all be done in real-time. We can give feedback to gaming systems within 300 milliseconds, which in turn helps to build a new breed of machine learning enabled tools. Clients can immediately react to the conclusions reached by our models and put them to work for the customer right away, creating a far more effective service.”

Many operators have previously taken a scattershot approach to data collection that has arguably resulted in information overload. In your opinion, what are the key bits of data they should be looking at?

“At Golden Whale, we believe in collecting data from the moment you first get to know your customer and follow them through the marketing funnel into your product and the game mechanic they’re engaging with. It’s only by having this complete picture that you can draw the most accurate conclusions. As you rightfully said, the challenging aspect is that there’s a potential information overload, as data storage is available and cheap. We live in a glorious age of abundance of real-time data when it comes to information – especially in the gaming industry – but it needs to be stored in a way that allows you to utilise it effectively.

“Our philosophy is “the more data, the merrier”, as we can then transform, and trim data based on individual use cases and the fundamental question that you want an answer on. More importantly, our algorithms enable you to find the right questions to ask – your “unknown unknowns”.”

Prior to the advancement of machine learning and process automation, what were the main pain points for companies when it came to data collection?

“Ironically – the abundance of data that is available. Our team has a background in biotech and medical computer science where it was previously very expensive to get good data to build models on, but in gaming, real-time data is digitally available for decades! Just five years ago, the challenge was storing and sorting all this data effectively so that it could be quickly moulded and applied, but this is getting easier by the day with advancements in technology and software. From 2015 onwards, many relevant new algorithms have surfaced that have brought us useful machine learning. This has shifted science from the academic space into the applied space and has quickly given us useful insight. And the perfect timing to continue building our products in this space.”

Your company is also involved in using data to assist businesses with game optimisation. How does this work in practice and have you identified any trends in terms of what users are looking for?

“I don’t want to spill the secret sauce here, but there are definitely trends and patterns. As we’ve worked directly with game developers, we’ve had results showing that certain betting patterns and the adoption of specific features are good indicators of future strong users. These findings can then be fed back into the game design itself, leading to the creation of games that inform the operator how to deal with customers in a more specific way – sensor games, as we call them. As a game developer, being in the position to feedback valuable information about clients to an operator makes things more interesting.

“There’s definitely a lot we can still learn from gaming data and I think the fact that Golden Whale is working on both fronts allows us to combine the knowledge of game operator and developer to benefit both parties.”

In terms of CRM, can you give us an example of how an online casino might use the data they’ve collected to automate a customer retention strategy?

“Time and resources are always in short supply in CRM, so one of the most important aspects we discuss with our customers is early VIP detection. We aim to provide – with a great deal of certainty – an accurate segmentation of their user base and what their future value will be. Armed with this information, CRM teams can better focus their resources and immediately go after the high-value customers that they think have future VIP potential.

“By consistently predicting future value with generation after generation of customer, operators can subsequently improve the long-term value of their player base. We are moving away from simple thresholds that are met after weeks or months to early predictions after just hours or days of gameplay.”

Does the ability to automate retention campaigns reduce some of the burden on CRM teams and allow the business to deploy their resources elsewhere?

“For sure. It’s already been done in other business areas and I’m sure it will be the same for CRM. The tools that are built out of machine learning applications are essentially assistants to help find the answer to complex questions. When the parameters or variables exceed 10-15 points, machines are very effective to find the best strategy, so this assistance will – with growing trust – lead to future automation. That said, I don’t think anybody in CRM has to be afraid of that. The machine will not replace the CRM operator, but rather support the operator who is confident in working with AI. The basic operation can be automated soon, but there will still be somebody steering the feeds of new content and ideas into the platform – and this will be the human role. Nobody is talking about someone manually creating receipts for online payments today. We’re not saying, “the bookkeeper lost his job.” No, there are jobs being built around these systems that give humans more time to interact with other humans and be creative.”