Talk data to me: how a more data-driven approach by the UKGC could mend industry ties

As we await the much-delayed white paper into the Gambling Act 2005 review, the scope for more investment in data capability within the UK Gambling Commission (UKGC) is one of the areas expected to be addressed.

The importance of data was highlighted by one of the former de facto gambling ministers, Chris Philp, in his speech at the GambleAware conference on 8 December 2021 where he touched on improving the quality of data that the regulator has access to. Philp also referred to the development of a data repository as well as data sharing as part of the single customer view solution.

Although just this June, the UKGC came under fire for its lack of knowledge on metrics during the DCMS Committee of the House of Lords inquiry into ‘What next for the National Lottery?’. During that session, UKGC CEO Andrew Rhodes was quizzed by chair Julian Knight MP on what metrics the regulator uses to decide whether the National Strategy to Reduce Gambling Harms works. When asked outright about whether there are any metrics, Rhodes responded: “I would need to write to you on exactly how we measure them.”

David Whyte, senior associate at Harris Hagan, agrees there is little to suggest that the regulator is making the most out of the data it already has, how it uses it and its internal metrics for assessment. On the flip side, the more information the regulator has, the greater the risk of its exposure to criticism if it fails to identify issues.

Melanie Ellis, partner at Northridge Law, echoes the argument for a more data-driven approach. “It seems clear that the direction of travel is for the Gambling Commission to gather more data from operators, use it in a better and faster way to identify issues and also to see whether their own approach is effective,” she comments.

Review time

Revisiting the topic of data again was the subject of a blog published in October by Simo Dragicevic, a member of the UKGC’s Digital Advisory Panel, examining the challenges and opportunities for the increased use of data to develop better gambling regulation.

It sparked industry chatter on LinkedIn, especially around the idea of a sunset clause like the one adopted by the alcohol industry in Scotland for the minimum pricing policy, which aims to save lives and reduce hospital admissions until Parliament votes in 2024 on whether it should continue. In his blog, Dragicevic explains that while he doesn’t advocate replicating this exact approach for the gambling sector, review periods for policy changes could prove useful to allow ineffective requirements to be removed or adapted.

Backing the theoretical idea of a sunset clause, Kirsty Caldwell, founder and director of Betsmart Consulting, believes it could be beneficial as multiple examples of regulatory requirements, once rolled out, have proven to have little positive impact on the consumer.

Using the example of the reality check requirement, Caldwell tells EGR Intel that significant technical work and roadmap-disruptive deadlines were mandated from operators to roll it out years ago, yet the use of the safer gambling control among customers has remained extremely low. However, there are some hurdles to overcome with this concept. “Theory aside, in practice I see some issues with using the sunset clause approach, the biggest one being the substantial amount of development work often required to ensure compliance with many new regulatory requirements,” says Caldwell. She also points out that with the alcohol industry being mainly physical rather than digital, rolling back a regulatory change might be easier to implement for that sector than for online gambling.

While Ellis can’t see the UKGC applying a sunset clause to reverse the credit card ban for example, it could be used to look at stake limits or the spin speed for online slots games. “So, they could start by taking a stricter approach than might turn out to be necessary, with the aim of reducing potential harm. They could then look at that later on and see from the data that perhaps there doesn’t need to be quite such a large delay between spins for online slots, or that a slightly higher minimum stake would actually still achieve the aims they are looking for, based on the data,” she notes.

Sharing his personal views with EGR Intel, Dragicevic agrees that applying a sunset clause to understand the contribution of individual slot game mechanics on harm would be beneficial. “If a fixed evaluation period existed, it could allow for different types of studies to be planned and conducted better, for example natural experiments, so a slot provider releases two versions of the same slot with different configurations, with one configured to be ‘less risky’. This could be one with a minimum spin cycle speed of 2.5 seconds versus the same slot set with two seconds, with the player behaviours triangulated with customer surveys to assess whether risk is affected by their play on the different slots,” he suggests.

Simo Dragicevic, a member of the UKGC’s Digital Advisory Panel

Any review period to allow for ineffective or disproportionate requirements to be removed or adapted from policy must adhere with the Gambling Act 2005. The act requires the UKGC to consult before specifying a licence condition and before issuing or revising a code of practice from which any formal requirement must derive. As such, before any change can be removed or adapted during the period, the regulator would be required to conduct a further consultation.

On this point, Whyte notes that this would be a complicated and confusing process and that there are better alternatives to be explored first. “Ineffective or disproportionate requirements could be easily avoided were the Commission to engage more widely with the industry, in particular its licensees, before it acts. Unfortunately, such engagement has decreased over time and feedback provided in response to consultations is often ignored or brushed aside with limited explanation.

“The result is that licensees are now of the view that it is highly unlikely that the Commission will listen, so there is reduced motivation to meaningfully engage. While there are indications that this may be improving, industry engagement is not where it should be and this is to the detriment of all stakeholders,” concludes Whyte.

A problem shared

Another aspect of data transparency is around the effectiveness of algorithms. In his blog, Dragicevic refers to gambling industry harm-reduction algorithms, which could be classified as those performing ‘safety’ functions in the draft EU law on building trustworthy AI, thereby warranting greater candidness and scrutiny.

Dragicevic, who also sits on the board for the Responsible Gambling Council, advocates the industry discloses information on the use of algorithms as the knowledge and methods are often not easy for regulators to access or interpret. However, there are pros and cons in doing so. He tells EGR Intel: “There has always been pressure not to disclose more due to competitive reasons on how algorithms and solutions are developed, especially for third-party solutions. There is also the risk that disclosure exposes gaps in sophistication and processes, and while this is a very good thing as it drives improvements, the industry will be concerned it puts further regulatory optics on executives.

“The gambling industry needs to weigh up the risk of greater disclosure versus the risk that regulators are forced to adopt more standardised approaches because the advanced approaches are too opaque and the benefits unclear.”

While transparency in relation to algorithms could drive greater acceptance around the benefits of their use as well as improving trust between the regulator and operators, there would be some hurdles to overcome first. For instance, the ability of data scientists to clearly explain how they work and ensure they have been understood correctly.

Caldwell explains: “Despite being a user of safer gambling algorithms throughout my career, I have yet to meet a data scientist who can fully explain – on a level which I can understand – exactly how the algo works.”

As referenced by Dragicevic earlier, another potential stumbling block is the risk of oversharing sensitive information with competitors. “For example, competitors may amend their thresholds and triggers to their advantage. This is not only unattractive from licensees’ perspectives but also for consumers,” points out Whyte. “Equally, consumers may seek to avoid snagging algorithms designed to protect them if there is too much transparency,” he notes. Yet Whyte adds a caveat that licensees are less likely to have concerns about disclosing information on how their algorithms operate if they were to understand why this information is being requested and how it will be used.

The other risk is that the UKGC could take what it thinks are the best approaches and hold all operators to those standards. Ellis argues that smaller operators could be severely disadvantaged by not having the capability to develop the same sophisticated algorithms as larger operators. A subsequent issue could arise if the regulator applies best practice expectations retrospectively.

“For example, in recent regulatory action against the operator of the Betfred brand, the Commission criticised them for not having controls in place to prevent large levels of high velocity spend by new customers. That’s not even a specific requirement of the LCCP now, let alone at the time that the relevant customers were playing, but it’s something that the Commission has picked up that some operators are doing now, and it seems to be applying that as an expectation across the board,” she comments.

Presenting evidence

The overarching benefit of the UK regulator adopting a data-driven approach would be to improve the UKGC’s relationship with the industry, for the greater good of the consumer, while being able to better defend itself from media and political attacks. “As operators have come to learn, it’s much easier to explain and defend your decisions if you are able to provide evidence that robust data has been gathered to drive them,” remarks Caldwell.

While some stakeholders raise concerns about regulator and industry co-operation on regulatory initiatives, Dragicevic wholeheartedly agrees that this cannot progress without such collaboration. He sees three areas in which operators can play a critical role in improving data-driven regulation in the UK. Firstly, by prioritising internal management and technical resources to help the regulator access the right data sets and to design efficient processes to integrate data on a regular basis, especially during any early test-and-learn phases. Secondly, working on standardising how risk data is reported at a high level to enable cross-operator and industry comparison and review. And finally, looking at how the industry can voluntarily disclose more information on its risk management governance, systems and processes. This will have multiple benefits, for example, educating key stakeholders in the work the industry is doing with data and algorithms as well as helping to share best practice and raise standards across suppliers and operators.

However, Northridge Law’s Ellis casts doubt on whether the regulator is ready or able to adopt a data-driven strategy right now. “They have talked for a couple of years about investing in data expertise, giving staff additional training and employing people with real knowledge and understanding of these types of issues. But I think it’s going to take them quite a while to get up to speed.”

Melanie Ellis, Northridge Law

Ellis highlights that there is no point gathering data if the capability to properly analyse it and draw conclusions from it is absent. “So, I think we’ll see a gradual increase in the data the Commission requests and what they’re doing with that information rather than any sudden changes,” she shares.

For a data-rich approach to work, Dragicevic concurs that the execution of it will be a challenge for the regulator to develop the internal capabilities to source, store and analyse the data, which will in turn require investment and time. “I expect that there will be a period of testing and learning, to enable the regulator to refine its approaches with industry support and to consult wider stakeholders,” he states.

While there are clearly a whole host of challenges the UKGC will face in attempting to adopt a more data-focused approach, the benefits are clear. Regular and reliable access to data could allow for more targeted assessments as well as bringing the regulator in line with the decision-making approach it already demands of its licensees. As LinkedIn executive chairman Jeff Weiner puts it: “Data really powers everything we do.”