Statistics and research release
Unlicensed gambling – Using data to identify unlicensed operators and estimate the scale of this market - October 2024
Using data to identify unlicensed operators and estimate the scale of the illegal gambling market.
Summary
Also published recently
This paper outlines how we are using data to improve our evidence base around the unlicensed gambling market. This work is contributing to our disruption activity of this market by automating and scaling up processes to identify unlicensed sites which were previously undertaken manually.
This paper also explains how we are using data to obtain insights into the overall scale and dynamics of the unlicensed market. We are taking an iterative approach to development of our use of data that can be implemented quickly and improved over time. As we expand this work, we anticipate further benefits, such as helping us evaluate any knock-on effects from regulatory changes and understanding how these influence consumers use of unlicensed gambling.
The first iteration of this methodology is subject to several assumptions and limitations and does not capture the whole online unlicensed market. We actively encourage operators and other stakeholders to consider how data they hold could be used to improve this model.
Details
Focus of this work
It is illegal for anyone to offer commercial gambling services to consumers in Great Britain without a licence from the Gambling Commission. This includes cases when no licence is held by the operator, as well as when the operator holds a licence in another jurisdiction, but operates in the UK without a license from the Commission. In this paper, we use the term unlicensed gambling to refer to any activity which falls into this category.
Although unlicensed gambling can also take place in land-based premises, the focus of this paper is the online market as this is where data has the greatest potential to help us make an impact.
Several stages of work have been undertaken to formulate our approach.
Overview of development stages
We undertook a scoping exercise of existing methods used to collect data to disrupt unlicensed gambling.
We reviewed a range of research by third parties and other regulators.
We spoke to several industry stakeholders, including hosting a workshop with experts from affiliate marketing, payments providers and internet technology firms.
As part of our consumer voice programme, we are conducting a multi-phased approach of quantitative and qualitative research to understand consumers awareness of the online unlicensed market, as well as how and why they access them.
Understanding consumer motivations
To help us identify the most useful potential sources of data, we created a typology of motivations for consumers to enter the online unlicensed market, and the channels through which they do so. These include those seeking better offers, avoiding know your customer (KYC) documentation checks, people who have been banned or blocked from licensed websites, and people who have self-excluded from licensed sites through GAMSTOP. The channels into the unlicensed market included affiliates, influencers, social media, and adjacent activities like crypto trading and gaming.
In practice, some of these motivations may overlap and will not always be mutually exclusive. We are also aware that some consumers could be using these sites intentionally, knowing that they are not licensed, while others could be using these sites without knowing they are not licensed by the Gambling Commission.
The data sources required to understand use of unlicensed websites by these different groups of consumers are likely to vary. We have decided to initially focus on specific areas of consumer motivations: people who have experienced gambling harms – especially those who are self-excluded; and consumers looking to avoid identity verification.
Our rationale for prioritising these groups was a combination of their vulnerability to harm and the potential size of this consumer group. As the work progresses, we will add additional areas of focus. This may or may not increase the number of sites that are identified.
Identifying unlicensed operators and estimating the scale of the online unlicensed gambling market
With a better understanding of why and how consumers access unlicensed gambling websites, we can identify ways in which we can use data to identify unlicensed websites and make estimates of their usage by GB consumers. We developed a methodology that combines web traffic data and gambling behaviour data to estimate the Gross Gambling Yield (GGY) of the online unlicensed market.
Obtaining an accurate estimate of the size of the online unlicensed market is, however, very challenging. Much activity is deliberately hidden – such as by Virtual Private Networks (VPNs). Nonetheless, useful data sources, such as the web traffic data, can give valuable insights.
An overview of the methodology is as follows.
Google search results to list of search terms
Results to search terms are monitored on a monthly basis. These search terms are designed around the unlicensed gambling motivations identified through industry engagement and our consumer research, as well as advice from our intelligence and enforcement teams. Additional desktop research is performed to identify terms used on affiliate pages to target particular groups of consumers, for example "not on GAMSTOP" for individuals who have self-excluded. A combination of Google Trends, and Similarweb's key word generator are then used to identify the most popular search terms by traffic. The top 5 pages of Google results for each search term are collected to ensure the majority of traffic is captured.
Identify affiliate pages or articles listing unlicensed sites
From these results, affiliate sites and/or articles are identified. These web pages recommend a list of gambling websites, often targeted at a specific consumer group, for example, “best UK casinos not on GAMSTOP”. This is done by checking for key words on the web pages and identifying the presence of outgoing affiliate links.
Extract links to unlicensed gambling sites and obtain web traffic data
The outgoing links to unlicensed sites from these affiliate pages and/or articles are then extracted. The unlicensed sites are examined to determine whether they are blocked to GB consumers. We know that some sites are blocked immediately upon opening, whereas some others are only revealed to be blocked when the user tries to register for an account. The current methodology is able to flag sites that are blocked immediately upon opening, but not sites that are blocked upon account registration. Web traffic and average visit duration data is then obtained for each of the unlicensed sites using Similarweb, which is a digital intelligence platform that allows access to estimated web traffic data.
Combine web traffic data with research data to estimate spend on the identified sites
The web traffic data is then combined with an estimate of average consumer spending behaviour to estimate the GGY associated with the identified sites. This has been taken from data from the Patterns of Play research. This is based on analysis of account level data from 7 major licensed operators, covering a period from July 2018 to June 2019. Information from 139,152 online gambling accounts provided insights into user behaviour, including an average GGY per minute of £0.32 for online slots1.
The outputs of this work are twofold. A dashboard of unlicensed operators ranked by their current usage by GB consumers. This can be used by our enforcement teams to help them prioritise and target their disruption activity. By delivering this through a data science approach, data can be delivered more frequently to give up to date insights into current activity in this market. The other output will be estimates of the likely scale of the unlicensed market for GB consumers.
Assumptions and limitations
This model can give us an indication of the trends in online unlicensed gambling activity. However, a number of assumptions are made with the current methodology and there are also inherent limitations to the accuracy of the model. A list of the assumptions and limitations are as follows.
Gambling behaviour on unlicensed sites is the same as on licensed sites
The Patterns of Play data used for Gross Gambling Yield (GGY) spend relates to behaviour on licensed gambling sites. There is anecdotal evidence from our Consumer Voice research into unlicensed gambling that people’s spending habits are different on unlicensed sites compared to licensed sites, as well as from research undertaken in other jurisdictions. For example, consumers may be more cautious about how much they spend if they know a site is not regulated in GB. Conversely some consumers may be looking to spend higher than average amounts if they are deliberately seeking to avoid Know Your Customer (KYC) checks. More research and data are required to inform and improve this assumption.
Research data used for our GGY estimate is from 2018 to 2019
The Patterns of Play research was based on data from 2018 to 2019 and there could have been changes in online player behaviour since that time. As we gain access to more up to date player level data, we will obtain an updated GGY per minute value to incorporate into our methodology.
GB traffic from consumers using a VPN is not captured
Some unlicensed sites do not allow people to access them from IP addresses based in Great Britain. Consumers can bypass this by using a Virtual Private Network (VPN) to mask the location they are accessing the site from. The model does not capture any traffic to unlicensed sites from GB consumers using a VPN.
GGY estimate is based on online slots play only
The first iteration of the model uses GGY data for online slots as we assume a significant proportion of unlicensed gambling activity is slots. As a result, the current methodology may not account for high spending consumers on other gambling activities, such as sports betting.
Unlicensed sites are included in the GGY estimate regardless average visit duration is
It is possible that sites with very short average visit durations could indicate visits where no money is spent. Similarly, very long average visit durations could indicate periods of inactivity where no money is spent. This is not accounted for in the current methodology.
Not all consumer motivations are currently included in our core search terms
We have prioritised collecting data based on certain consumer motivations for accessing unlicensed operators. As we explore other motivations, we may identify additional operators which are not currently captured by our existing approach.
Next steps
Our work to date has allowed us to:
- scale up our online detection of unlicensed sites, and provide a more frequent, monthly feed of data for our enforcement and disruption work
- start to build-up estimates of scale of the unlicensed market – which can continue to be tracked over time to understand ebbs and flows in its size and dynamics.
This work, however, is just a starting point. We have ambitions to continue to grow and improve our use of data in relation to this topic. To improve the current methodology, we will:
- add additional search term themes that reflect consumer motivations not captured within our phase one model
- incorporate questions on the GSGB to better understand the behaviour of consumers who use these unlicensed sites that will help to address the assumptions made in the gross gambling yield (GGY) estimate.
Findings from our methodology will be published in late spring 2025. By then we will have accumulated 12 months of data and had an opportunity to further develop the methodology by starting to address some of the assumptions mentioned in this paper. We will also apply this accuracy of our estimate under current assumptions.
We are also interested in exploring other channels into the market, including the following.
Social media
Our initial methodology has focused on search engine results. We know that promotion of unlicensed gambling sites also takes place through other social media platforms.
Encrypted communication apps
Encrypted communication platforms, such as WhatsApp and Telegram, may be used to promote unlicenced gambling sites. Identification of this content is challenging because of the high levels of encryption. We are interested in stakeholders’ views on how data can be collected from these sources.
Adjacent activities
Gambling slots streaming sites, such as Twitch and Kick, and crypto trading platforms are also locations where gambling sites without GB licences may be promoted. We want to understand if this is generating further traffic from consumers with a primary interest in video gaming or trading.
Open banking
Test purchasing could be used on unlicensed sites to identify merchant codes. These could be used as part of analysis of open-banking data to obtain further insights.
Notes
How you can help
This work serves as a first step in understanding the scale of the unlicensed market. By taking an iterative approach, where we are transparent with our methodology and reasoning and learn through our experience, we will benefit from a more collaborative approach, taking input from all stakeholders.
We know that other stakeholders, including licensed operators, may have data that will be useful for improving our assumptions or recommending new methods of detection. Tackling the unlicensed market is a shared goal, and we encourage any feedback for ways we can improve our methodology.
If you are aware of any data that could be shared with us to address any of our current assumptions, or any information on other motivations or channels for the unlicensed market, we would like to hear from you. Please use the feedback form.
Data and downloads
There are no files for this release.
Feedback
We are always keen to hear how these statistics are used and would welcome your views on this publication.