Experts Table with Betbazar: Gambling needs

Experts Table: Gambling needs a sanity check on data discipline & business intelligence‍

SBC: Are businesses really making the most of data? If not, why? What’s stopping them?

Etienne Azzopardi: I believe one only has to look at the term “data mining” to realise how important it is to the iGaming industry; we’re likening a widely-available digital commodity to a precious metal – and for good reason. That being said, while businesses probably don’t need any convincing in regards to the value of data, the fact remains that not everyone is currently making the most of it.

There are, of course, a number of reasons why this is the case, one of which is that vast swathes of data are perhaps siloed across different systems and departments, meaning companies are potentially unaware of the insights they’re missing out on. A lack of expertise and data-literacy also play a role in why businesses are not identifying data initiatives and the potential value and cost-effectiveness of data integration between systems. It’s not unheard of for companies to be caught out by skyrocketing BI costs and not have a lot to show for it in terms of business value.

Max Sevostianov: There’s an old saying from the 1800s that goes: “who owns the information, he owns the world.” While that’s still just as true today, I think we now need to update the key word from “information” to “data”. Certainly when we look at the more successful companies in our industry, they all now collect and analyse data to improve their business models and products.

For those that don’t currently do this, there are a couple of factors that might be preventing them – namely cost and not knowing where to start. We hear a lot about data scraping – which is when a company pulls their data from other websites – but this is a scattershot approach and is always a bit risky. Usually it’s better to invest in official data to achieve optimal results for your business.

Allan Stone:  From what we’ve seen when working with gambling businesses, they aren’t making the most of data simply because they don’t have streamlined access to information.  In many cases, they are sitting on a treasure trove of data, but they don’t have the ability to take that data from specific silos and move it into a format which is actionable. This is one of the major pain points that our partners face.  

We work alongside our partners to help them understand where their data lives, how they can access it, and then how they can make that data actionable by marrying it with other sets of data that exist within their platform.

Thomas Kolbabek: We get a very diverse impression about what’s going on with data handling, analytics and data exploitation from the companies that we work with in the iGaming industry.  

We come across all kinds of scenarios, from structures that struggle to even consolidate data or handle the amount they generate, to products backed by quite sophisticated models which are still very slow in the adaptation and generation of results.

What’s stopping most of these companies from performing better is the quickly widening gap between what can be done and what these companies actually do to modify these structures.  

This is why we not only focus on building models for the gaming world but also develop a complete data handling and analytics environment with firm foundations. This allows us to interact with our partners exactly at the level they need to achieve significant growth.

Karl-Jorit Hausdorf: At Fujitsu, we have conducted a joint research with Freeform Dynamics, the analyst company, which looks specifically at this question. When you say making the most out of data, that is an incredibly broad question, but for me there is one clear answer: research. The research that we have conducted suggests that most companies are not making the most out of their data.  

Across Europe, we have seen a split between companies that are data-starved, data-sustained, data-empowered, and data-driven. What we found is that the majority of companies sit within the data-sustained and data-empowered categories.  

If we look at what is stopping them from making the most out of their data, there can be lots of different things. We categorise the ability to be data mature as an enterprise with four criteria, the first of which is ‘do you have the culture and mindset?’.  Sometimes, data can be kept in a silo, but it should really be shared across an entire organisation, especially if you want to become a data-driven business.  

The second criterion is user experience. In the more mature organisations, they may have a team of data scientists. But if it’s taking that team several months to get the right data, they’re undoubtedly going to go on to the next job. So it is really important that the user experience to consume data is streamlined and accessible. Ideally, requirements are set upfront which allows them to analyse data in real-time. A lot of companies elect to look at data purely from snapshots, but in order to get a clear picture and identify trends, that data needs to be collected and stored every day. That’s one of the most important things, where companies don’t currently think that temporary data is important.

If we look at data management, we have to first look at what it means to be a data collector. That also means that you clean up your data – this doesn’t necessarily mean that you delete the data you have, but more that you can determine what data is business critical . Only 20% of the data that you store tends to be business critical, so it’s important to distinguish that information. This becomes even more important when you begin to look at technologies such as AI, because you cannot feed AI incorrect data. You need to manage that data up front and have clear ownership of contingency management.

System infrastructure is also really important. Let’s say you are about to invest into AI features, they need to be available at the right speeds for data input and data output as required. A lot of companies don’t think about this upfront. It’s not a given for many companies that are ready to deploy data-driven technology by default.

Andrea McGeachin: Some businesses are effectively leveraging data by investing in decision science rather than mere data analysis. The crux lies in Data Decisioning, which unlocks the true potential. By amalgamating diverse data sources, we can unveil insights unattainable when examining a single stream of data. Despite the scarcity of companies prioritising this aspect, the trend is on the rise.

How can businesses identify the right/best data to work with?

Etienne Azzopardi: I think the first step for any organisation is to define their data strategy based on what their business is trying to achieve and align the data being targeting accordingly. With this strategy in place, they can begin conducting a stocktake of each data source they possess and identify the information that carries particular relevance, value, or quality in relation to their goals.

Of course, any data analytics should always be carried out with the input of business stakeholders, and this should result in reports and visualisations which – through an iterative process based on feedback – can then be refined and elaborated on. It is only via this ongoing cycle of analysis and development that companies can really identify the data that is most relevant to them and how it can be harnessed to add value to their business going forward.

Max Sevostianov:  I think this generally comes down to identifying what areas of your business you need to improve and what goals you’re hoping to achieve. Usually, a company will want to make a detailed analysis of a specific pattern or KPI number and it’s just a matter of determining which data sources will provide the most relevant insights for that topic.

If it’s a question of how customers are interacting with a particular product and how that product is performing, then focusing on internal data is key. However, if we’re talking about optimising your trading process and delivering odds that have the biggest benefit on your margins, then purchasing official data from external sources will be the thing that really enables you to improve your product and deliver the best up-time possible.  

Allan Stone: I think that comes down to understanding what the specific goals for the business are over the month, quarter, year and then down to the individual daily and campaign performance level.

When we look at data, we’re looking at it very specifically from a campaign performance perspective; we’re advising companies on what data they should be looking at and helping them to understand how to set goals for particular campaigns. We then look at how those goals align with the overall business objectives for that timeframe and act accordingly.

Thomas Kolbabek: It depends largely on the questions you want to ask from the system. The general rule would be to generate as much data as possible and then start a selection process as soon as possible, to filter down to the needs of your specific setups.

Karl-Jorit Hausdorf: That all depends on the use case. The use case is just 10% of the effort of the whole journey of making a data-driven solution happen. In the past, companies were more inclined to clean up all their data, invest in big data warehouses and start projects that would never end. But now, it’s more important to identify the use case, which should come from the questions ‘what are the users supposed to do in their daily business? Which data to the users need at hand? Does it need to be real-time?’

If those users get the right data in the required form, you can enhance it even further with analytics, predictive machine learning algorithms, etc. What would be their benefit? And that would form the third option, which is the use case together with a business case. The business case normally looks at the question ‘can we make those users more efficient?’.  AI is like the biggest hope in achieving that; it can help businesses to become more efficient. Since we are moving into generative AI, it’s not just about efficiency, but also the actual output, the performance, the quality of their work that can be tremendously enhanced by using data matching with generative AI.

Andrea McGeachin: Every piece of data reveals trends and areas for enhancement. My top recommendation for companies is to thoroughly analyse all data, leveraging skilled employees in conjunction with advanced software. The crucial questions to ask are: What insights does your data yield? Which findings are actionable and beneficial for your business operations?

SBC: How important is identifying the right data to unlock its full potential?

Etienne Azzopardi: As touched upon in my previous answer, this is where it becomes critically important for data and commercial teams to be properly aligned on their desired outcomes. An expansive data project meant to identify the split between right and left-handed customers might be important if you’re producing computer mice or hand-held tools, but not so much if you’re a bakery – unless a market suddenly develops for left-handed bread, that is!

As such, data initiatives should always be primarily driven by the problems that the business is trying to solve, such as building a better product, assessing trends, and identifying gaps in the market where there are opportunities. Essentially, analysing data purely for the sake of it is largely unproductive and companies should instead set out with a goal they want to achieve and conduct their research accordingly.

Max Sevostianov:  It’s vitally important if you truly want to improve your product and optimise the service that you provide to customers. As I said in my answer above, the first step is always identifying the area that you want to focus on and tracking down the most relevant data.

When it comes to analysing the numbers and how a specific product performs, you have to remember that we’re not talking about getting one-on-one feedback here – we’re potentially reviewing information from millions of active users and drawing the right conclusions from your data can lead to much faster improvements. Likewise, when you invest in a new product, identifying the right data can help you better understand how it will be received by your customers and what the potential ROI might be.

Allan Stone:  It’s key, right? If you’re not looking at the right data, you’re not understanding the full picture. Quite often, when we work with partners, we focus on understanding the data that they have at their fingertips, where they might be able to look deeper, what they’re not looking at and where they can improve. If you’re not looking at the right data, you will face much more difficulty in making the right decisions.  

We have found that a lot of operators don’t seem to be getting the full picture from their data, they’re only seeing a small insight into what that data can achieve. What this means is that they are making decisions which aren’t fully substantiated; we help them realise their full potential.  

Thomas Kolbabek: At the moment it is more important to not miss out on important data points. We run fully automated filtering and selection processes so once selections are made, it’s easy to trim down on data volume if necessary.

Karl-Jorit Hausdorf: I think this is really important, if not the most important thing to consider. Actually, the way we are talking about data, I am a big fan about facts. Data is a type of knowledge that is being stored. You have lots of data sets that could be equally right and wrong at the same time. Companies may load that data into whatever system that they might be using, and it may include data that is wrong.  

Knowledge management therefore becomes even more important than then data. I would identify where my knowledge is first, and then see which data sources are delivering that kind of knowledge. If you if you don’t deliver the right quality, you won’t have the right values which will then impact your business case.

Andrea McGeachin: Any data from your system is useful but it’s the output that counts. By utilising basic data, we managed to dissect our team’s productivity and pinpoint areas of strength and weakness. When measuring AML, it’s about more than just recognising patterns – it involves analysing extensive transaction data to shape a robust AML strategy. However, data collection should always respect players’ privacy boundaries.

This analytical process often reveals unexpected insights, like diverse business models in specific regions. While data is pivotal, the ability to interpret data science plays a vital role in maximising returns on investment.

SBC: In which areas can data have the greatest impact? How can a business determine where data can drive the most improvements?

Etienne Azzopardi: Generally speaking, so long as an organisation has a clear objective in mind, or a problem that it’s trying to solve, data can play a significant role in improving performance in most areas of a business. I believe there’s a lot to be gained by analysing data on internal factors such as the performance of teams and the process of product development, as this can then be used to make relatively straightforward improvements that can have a tangible impact on business efficiency.

Again, for organisations that generate huge volumes of data – for example, the transactional data accumulated by online payment processors – it’s important to set a data strategy and revisit it as often as necessary to ensure its continued relevance.  With the right strategy and underlying systems in place, it’s not uncommon to unearth correlations that you may not have been looking for, but nevertheless have tangible commercial benefit to your business.

Max Sevostianov:  Be it internally or externally, data really can have a massive impact on all facets of your business. Whether you’re trying to improve an in-house workflow procedure or analysing your product offering to determine how certain user experiences could be optimised, all of the information you need will be contained in the data – you just have to know how to read it.

In terms of identifying where that data will drive the biggest improvements for your business, I think you just have to approach it from a problem-oriented point of view. In what areas are you not performing to expectations and what are the main reasons for that? Once you’ve done that, you can deep-dive into all the available data you have on that specific topic and identify how things might be tweaked to deliver better results in the future.

Allan Stone: It all goes back to understanding what those goals and objectives are. When we work with our clients, much of the focus is on acquisition campaigns. So, quite often, the data that we need to be looking at is centred around acquisition costs. But following on from that, we then look at what those players are actually doing post-acquisition and whether their behaviours influence whether those players will be profitable in the long run.  

When we analyse that information, there are a few different areas that we like to focus on. The first, and this might sound simple, but what is the point of analysing this information? We want to understand how best to create player acquisition campaigns – how much does it cost you to acquire that player? Are there segments of player acquisition that might be more expensive than others? Sometimes, even if that initial player acquisition cost is more expensive, the quality of those players is actually higher.  

In some instances, a lot of operators are pushing towards a customer acquisition goal. Achieving that goal is great, but you risk the quality of those players not being as high as they should be. When we’re advising operators, we want them to consider their acquisition targets – they should be focusing the long-term value of their players, rather than just the sheer quantity of bettors on their site.  

It is really about understanding how to be flexible with your campaigns and understanding how you’re going to effectively use that data to drive your end business goals.

Thomas Kolbabek: There is most likely significant room for improvement in every situation where there are multidimensional and/or non-linear problems to be solved as quickly as possible.

Improvement can also usually be achieved where there are large amounts of complex datasets to be searched for high-value optimization potential. Machines are by far better at dealing with these tasks.

Karl-Jorit Hausdorf: Fundamentally, this has also changed a lot through generative AI. Previously, there would be a lot of processes and you would perhaps consider automation to optimise your business. Companies would look at data visualisation or standard analytics just to get an idea of how your business works with a standard set of KPIs.  

The next step is going to be generative AI, where it’s more about how you are able to generate new data with the data you had previously collected, and how that information can then be used to provide more insights for decision-makers and consumers. That’s where generative AI will make a massive difference in the future.

Andrea McGeachin: The first thing that comes to mind here is how data can hugely impact AML, fraud monitoring and reporting. Enhanced data availability allows companies to detect suspicious activities promptly.

Furthermore, data aids in recognising player behaviours, guiding tailored interactions and targeted marketing strategies. This personalised approach fosters customer loyalty.

Data also influences business models, enabling smart decision-making for profitability without compromising fairness for customers. Tailoring onboarding processes based on individual player needs through data-driven insights enhances customer experience.

SBC: How can these insights then be turned into actions that ultimately drive improvements?

Etienne Azzopardi: In order for this to happen, the organisation must first be ready to adapt to a data-minded culture. It’s not uncommon for certain insights to be undervalued by industry veterans who trust their own experience more than what the data is telling them. While there are, of course, certain areas where expertise prevails, it’s important for any business to identify what’s “out of bounds” and what’s open for experimentation when defining their data strategy.

Once data insights have been gathered, the fastest way to drive improvement is to translate them into actionable tasks such as making the changes to the product that are required based on what the data is telling you. In this example, the product team should then be able to act on the insights in a timely manner to deliver the optimum results, while also understanding what’s possible and what’s not in terms of value.

Max Sevostianov:  I think this is a relatively straightforward process once you have the relevant data in front of you. You simply need to analyse all of the available information for patterns and trends, make an assumption as to why these might be present and create and test a hypothesis for how they could be improved upon in future.

Whether this results in you optimising a certain product feature, enhancing your security measures or personalising your content for specific markets, you just have to try your hypothesis, analyse how the data changes as a result and constantly re-test and refine this process until you get the results you want.  Just as I said above, the use of AI – particularly programmes that can automate improvements in real-time – make this testing loop much more efficient.

Allan Stone: What is really important is that we understand the different metrics that operators are using to determine success and failure of their campaigns. We then put those metrics in place so everyone from the CMO to the campaign managers can all operate off the same sets of data and make synergistic decisions. That way, there is no conflicting interests.

We’ve seen that sometimes in the market, you might have a CMO who is expecting aggressive growth in the number if FTDs. But then you might have someone on a campaign level whose performance is centred around the return on investment from individual campaigns. Those two goals could, theoretically, be at odds with one another – especially if the team doesn’t have access to the same sets of data. As you can imagine, that can then create a misalignment in what the business is trying to achieve.  

It all comes down to how you create an environment where everyone in the organisation understands that data, what it is saying and how you use that data to make sure everyone is rowing in the right direction.

Thomas Kolbabek: We’re developing next-gen operations infrastructure in which large data spaces are generated for experimentation, and to connect these results to actionable items on the end-customer side.  

Human ideas and ingenuity are more important than ever at this point because that’s how we generate as many relevant items to optimize as possible and to cope with the steeper learning curve that comes as a result of machine-driven optimisation.

Karl-Jorit Hausdorf: I think it comes down to the KPIs you set up from a business perspective. Let’s say you want to improve the process of hiring people, you would need to set KPIs, and then find the right data to underpin this KPIs. Then, ideally, you would change your processes from a business perspective to increase those KPIs, whatever they are. In HR, for example, it might be that you want to be hiring more talent from the market. So we need to improve that process of bringing in talent.  

You should look at data of how many people apply, how many respond to requests quickly, how many you want to interview, how you store that feedback from interviews. These KPIs will then form the required input for data as an exercise.

Andrea McGeachin: These insights can be turned into actions through commitment and management training. Initially, it is crucial to listen to your data team and determine the essential insights versus the ones that are merely interesting.  

During this process, it is vital to ensure that multiple managers and leaders are involved in the final decision-making. Senior leaders can assist in identifying overlooked areas and foreseeing the impacts of decisions on other departments.