Big Data Primer: A Q&A with Marc Wood, Gambler Analytics

Go back

When it comes to understanding big data and the impact it will have on the future of the online gambling industry, there is possibly no better person to speak to than Marc Wood. Now the chief executive of his own company Gambler Analytics, Wood has been involved in the sector for over 15 years – pretty much from the industry’s inception – working first at payments firm DataCash before moving on to Wagerworks (now IGT in Europe) and then at credit-check and identity-verification business Experian where he was head of gaming. Now with Gambler Analytics – which has signed on as another GamCrowd partner firm – he works with a number of leading-edge big data, analytics and blockchain companies. He is one of those figures who influenced GamCrowd to think about doing a ‘data month’. Big data is the science behind how and why players and consumer behave the way they do and once a business understands that – in real-time – then a world of potential opens up in terms of player metrics. Or as Wood says, affecting his best Michael Caine impression, it will ‘blow the doors off the industry’.

GamCrowd: Do you think the industry yet has a good understanding of what big data can do or means for the industry long-term?
Marc Wood: I think first off its important to define what we are talking about, is big data just volume or a set of new technologies to handle it? The term is used interchangeably but honestly it’s both. As for the industry understanding what big data (in both definitions) can do is yet to be proven. This is not just for the online gaming industry but most others. There are very few companies that have really mastered it, Google, Facebook and Amazon spring to mind, but even they are still learning. My personal opinion is that the use of big data enabled techniques in an operational, real-time way, will ‘blow the doors off’ historical gaming industry metrics we are all used to.

GamCrowd: Where has big data had the most impact in the industry to date?
Marc Wood: I think that the real material impact of big data is yet to come. There are some interesting experiments or proof of concepts and cool companies out there but along with most other industries the operationalization of big data-enabled processes its still in the research and how-do-we-do-that phase. I think the biggest challenge for most gaming operators is that their core legacy gaming platforms were never designed or built with big data technologies in mind, so as ever integration of something new is a rock in the road.

GamCrowd: What is the technology behind big data? How recent is the technology?
Marc Wood: Big data in a technology sense is a suite of software languages tools such as Scala, Python, R and Java for programming in Hadoop and Spark. Some of these have been around for over 10 years in their own right but most are evolutions of previous technologies so difficult to pin down a real birthday. An interesting read is the history of Hadoop to illustrate this.

GamCrowd: What will be the next big areas for the utilisations of big data?
Marc Wood: Real time player analytics for player retention for sure

GamCrowd: Can you explain how big data and the concept of machine learning interact? And what do you see as the opportunities?
Marc Wood: Machine learning is a type of artificial intelligence (AI) program that provides computers with the ability to learn without being explicitly re-programmed. As with all AI programs the more data and inputs that can be provided the faster they learn. In the context of online gambling if building models that can predict the behaviour of, let’s say, a poker player (which totally possible but still rare today) that initial model will always be a first stab and the current approach will be to keep refining on a test-and-learn basis and requires constant rebuilding of models in order to be able to better predict the future. So by using AI machine learning models will mean that the models will keep on improving themselves without the pain, hassle, time and cost of rewriting and re-building potentially hundreds of models. To illustrate this there are 10 different Texas Hold ‘Em player types, and therefore models to be built, and that’s just one game and just poker. Expand this across poker, casino, bingo and sportsbook and today it’s an impossible task but with machine learning it becomes a reality.