Home > Casino & games > Q&A: Rank’s Jan Teichmann on the applicance of data science

Q&A: Rank’s Jan Teichmann on the applicance of data science

| By | Reading Time: 5 minutes
Rank's head of data innovation Jan Teichmann looks at how its data science team's findings are influencing areas of the business from promotional activity to responsible gaming.

Rank started building out its data science team on the back of a major overhaul of its legacy back end systems and a move to a new digital gaming platform, and as its head of data innovation Jan Teichmann explains, their findings are already influencing diverse areas of the business from promotional activity to responsible gaming.

iGaming Business: What was the motivation behind Rank’s decision to set up a data science team?
Jan Teichmann: There was a decision to grow the business through innovation and investment in quite a lot of different areas. Rank made a big investment into its data architecture, replacing a lot of back end systems and legacy systems and switching to an entirely new digital gaming platform. It also decided to bring a lot of new roles into the business, with data science being one of these.

iGB: What type of backgrounds do the members of your team have?
JT: The backgrounds are extremely diverse and that is mainly because the requirements the business has in terms of data are very diverse. We all have an academic background and every one of us has a PhD, although the fields are very different — we have an astrophysicist and one of our data scientists comes from cancer research and genetics. We also have people with more typical mathematics and statistics backgrounds; it is a very diverse bunch of people.

iGB: What were the most surprising findings to emerge from this level of granular data analysis?
JT: Some of the more important learnings we’ve provided to the business were around customer value and the impact of promotional activity on customer value. We helped the business understand how to mitigate the risk of bonusing. We also learned a lot about how players interact with our products, the kind of patterns our players have and the journeys players take. We were also able to convince the business that there is a still a place for bingo. There was a tendency for a lot of conversations to be very slots-focused but our work showed there certainly is still room for bingo and that was a bit surprising for some because bingo was seen as a dying product segment.

iGB: The data science team’s creation was mentioned in Rank’s annual report as something that was expected to improve customer insight, customer yields and marketing efficiency, as well as customer experience. Can the resource allocated to data science be directly correlated with reduced spending in other areas or does its success depend on increased revenues?
JT: The data science team is part of Rank’s investment into its future and it measures that success in a lot of different ways. It is not about a reduction in spend but a packaging of budgets in areas where Rank sees it is sensible and can provide the highest return.
iGB: What have the main deliverables to the bottom line been (in terms of yield, etc.)?
JT: We do not own the execution channels so we do not even try to directly attribute any uplift we produce in data science to data science itself. We certainly know the kind of uplift customer operations as a whole —  the combined efforts of data science, CRM, the website optimisation team, the outbound team calling customers — is but we have never invested time in building an attribution model. We’d rather spend our time building customer-facing models and trying to improve the customer experience rather than fighting over who owns which pound. For example, the CRM team used our churn intervention model to help lower customer churn and they have seen double-digit returns from that. How much is down to the data science model highlighting the customers they should target and how much of that is down to them choosing the right way of targeting them, in the right moment, on the right channel, with the right message? It is quite difficult to attribute how much of that comes from the data science effort and how much from the CRM effort.

iGB: Is this one thing stopping other gaming firms investing so heavily, that it’s more difficult to assess the return on investment?
JT: Probably, because you always have the problem that you have to double count everything or you need to have a very sophisticated way of splitting revenue uplift between the different teams involved. The projects we are running are all joint efforts between different teams or between retail and digital.

iGB: How does data science help Rank improve cross-channel conversion?
JT: So far we have contributed a microsegmentation and a propensity model, which basically helps create a cross-channel strategy. We have a lot more ambitions, particularly around personalisation and microsegmentation of our customers, to help to create the right customer experience for each customer. Our cross-channel customers are very important customers for our business but they are different to the otherwise more traditional digital customers so we need to cater to that.

iGB: Given Rank was one of the first movers in data science in gaming, what other industries that are perhaps further into the journey did you look to for inspiration and lessons?
JT: Ecommerce is an industry we talk a lot about, in particular Asos is a business we reference a lot within customer operations, particularly with regard to our cloud ambitions. Rank has a strategy to be a cloud-first operator in terms of how our data is stored and how we work with that data. The data science team is making a big ongoing effort to move more of our work streams into the cloud and migrate our personalisation models into the cloud.

iGB: Were there things Rank was already doing, perhaps in terms of fraud or responsible gambling, that translated over to data science?
JT: Data science has completely transformed Rank’s responsible gambling efforts. What we had done before was we had this kind of customer ‘lookalike’ in that you take someone who is self-excluded or has developed a problem with gambling and you look for customers who look similar to this customer, but we were comparing customers when a problem had already surfaced. That is kind of too late; you want to intervene much earlier than that, so we focus a lot on trying to understand the customer journey and the change points and offer soft interventions much earlier, and maybe prevention if possible.

iGB: So in regard to the Gambling Commission having said gambling firms should be using data to understand problem gambling as well as to market products, would you say Rank is already doing this?
JT: Certainly. Rank’s ambition is to go way beyond what is currently required simply from a regulatory perspective and we have worked very closely with the compliance team and we try to lead in that RG field rather than just play catch-up with the latest changes in regulations.

iGB: Is there potential for data science innovation to work against gambling companies, for example, by creating products that empower gamblers and improve their strategies, perhaps most relevant to sports betting or poker?
JT: The data science answer for that is that no one can beat mathematics. But there is certainly some truth in the idea that there has been a big liberalisation of data, and machine learning is slowly becoming more widely available as a service, for example, Microsoft’s Cognitive APIs, so now everyone who can pay for it can have sophisticated state-of-the-art machine learning. But most customers play for entertainment — there are skill-based games like poker but there the casino takes a cut of the pot and that wouldn’t change if one of the players had access to better tools to play better poker. Where there will probably be some friction is that the liberalisation of data will probably start to work against the big gambling companies because it is more likely that one of the small digital gambling companies will use this new level playing field to challenge the big operators.

iGB: How to do you expect Rank’s data science team to evolve in future?
JT: We will keep on growing the team as we grow our success. We now have a team of 11 people and we have already outgrown the desk space that was allocated to us in our new office within two months. Our competitors better watch this space because we are just getting started.

Related articles: Rank Group sees pre-tax profits fall despite digital surge
Rank hands non-executive director role to Thursby
Can big data help the industry and regulators address problem gambling? (Paywall)
Data compliance: CSR-commercial balancing act required (paywall)

Subscribe to the iGaming newsletter