Game design
July 7, 2022
- Adam Rivers Director, Economics KPMG
Digital game design and the changing regulatory landscape
Exposure of potentially harmful digital design practices in gambling will place the industry under scrutiny, argue Adam Rivers and Jurate Markeviciute. This article was finalised and received prior to the Behavioural Insight’s Team publication on choice architecture and gambling – which only echoes the need for the sector to understand these issues and act accordingly.
Regulators around the world continue to raise their concerns over digital practices and how firms’ use of data is impacting outcomes on end consumers. As part of that, we’re seeing proposed changes in legislation in the UK and EU that would strengthen regulators’ powers in both consumer protection and digital markets.
In the UK, the Competition and Markets Authority (CMA) – a body that has recently gained significantly enhanced powers in relation to consumer protection – is increasingly concerned about the impact of certain online practices on consumers. In addition to its cross-regulator work focusing on algorithms, it also published two papers discussing how digital design such as algorithms used to frame choices can influence consumers’ decisions. Its concern is that these practices can potentially lead to harm.
How can digital design harm consumers and competition?
In online settings, businesses design the environment in which customers interact with the website and/or app and make consumer choices – e.g. reviewing racecards, browsing a live casino lobby or playing an online slot game. The way in which these options are presented is often known as the ‘online choice architecture’ (OCA).
In this environment, regulators are concerned that businesses may also use various, and harmful, OCA practices. Specifically, they worry that these practices, designed either deliberately or unintentionally, might negatively affect consumer choice, leading consumers to spend more, receive poor value services, or search less for alternatives.
Harmful OCA practices may persist even in competitive markets due to low OCA awareness (and their effectiveness in influencing consumers even when recognised), potential profitability of OCA practices and/or certain features of a market.
Setting the scene: behavioural biases
Until relatively recently, regulators studied consumer outcomes under the assumption that consumers made rational decisions. However, the emergence of behavioural economics in recent years has shifted that thinking. In short, consumers are not always rational. Preferences are not always consistent, with us often relying on heuristics to short-cut decision making. Fortunately, many of these behavioural biases are systematic – meaning they can be anticipated and understood. This leads to some concerns, namely that businesses may be able to exploit these biases for commercial gain, and that digital markets have the potential to exacerbate them. However, regulators are also increasingly able to use evidence of these biases to intervene in markets more effectively.
Emerging thinking: online choice architecture
Online Choice Architecture (OCA) is the design of the online environment where users interact with businesses. This design affects our decision making and actions when we browse, compare, play and shop online. In April 2022, the UK Competition and Markets Authority (CMA) published a summary paper and accompanying evidence base in relation to OCA. It is this paper that we summarise here and apply to the sector.
It is worth noting that the gambling sector is featured prominently throughout the CMA’s paper, both in the context of previous investigations it has conducted and the volume of academic literature available that demonstrates the use of OCA practices in the sector. Our view is that the sector should therefore be braced for further investigation in this area going forward.
Three categories of practice
The CMA outlines a possible taxonomy of 21 OCA practices that could be used by businesses, as well as consumer and competition authorities, to help recognise, categorise and explain the impact of practices. These are broadly categorised into three types (although these are often interlinked and can be grouped in different ways):
- a) Choice structure is how choices are presented to consumers.
- b) Choice information is the information provided to consumers when presenting choices.
- c) Choice pressure is how consumers’ choices may be indirectly influenced.
The tables overleaf provide the full list of the practices considered in each category, together with ratings of the strength of existing academic evidence underlying them. These evidence strength ratings are an assessment of the extent and quality of available academic research relating to each OCA practice.
The majority of these practices can be, and often are, used beneficially, or are harmful only in certain circumstances. However, as suggested by the academic literature, some practices are almost always harmful (marked with “%” in the tables overleaf.
Choice structure overview
The CMA finds that there is strong evidence that choice structure practices alter consumer behaviour. Depending on how they are deployed, practices can have both positive and negative impacts on consumer choice. For example, well-designed ranking and defaults can assist in making decisions more efficiently, but these practices can also exploit consumers and lead to them choosing worse options.
However, half of the choice structure practices set out by the CMA are found to almost always harm the consumer, with potential outcomes including more costly and/or inferior decisions, including purchasing unwanted products. Some of these practices or iterations of them (such as particular variations of forced outcomes) are automatically considered to lead to negative outcomes and are already banned in the UK and other jurisdictions under consumer protection regulations. Mitigations available include promoting active, meaningful choices to consumers and recognising where friction is required or can be removed in ways that are shown to benefit consumers.
Table 1 below shows the CMA’s summary of OCA practices relating to choice structure and a high level description. We then move to discuss three of these in detail in relation to the gambling sector.
Table 1: Choice structure examples
OCA practice | Description | Evidence |
Defaults | The choice architect applies a predefined setting that the consumer must take active steps to change | **** |
Ranking | The choice architect displays the order of options in a particular way | *** |
Partitioned pricing | The choice architect presents individual price components without sharing the total or estimated total costs with the consumer | *** |
Bundling | The choice architect groups two or more products and/or services in a single ‘package’ at a special price | *** |
Choice overload and decoys % | The choice architect provides too many options to compare. The choice architect adds an option to the choice set to make the other option(s) look more attractive to the consumer | *** |
Sensory manipulation % | The choice architect employs visual, aural and tactile features to steer consumers towards certain options | *** |
Sludge % | The choice architect creates excessive or unjustified friction that makes it difficult for consumers to get what they want or to do as they wish | *** |
Dark nudge % | The choice architect makes it easy or removes friction for consumers to make inadvertent or ill-considered decisions | *** |
Virtual currencies in gaming | The choice architect creates elements of a virtual currency to be used as a substitute for the ‘real-world’ currency | ** |
Forced outcomes % | The choice architect changes the outcome without giving consumers a choice | ** |
Choice structure – betting and gaming examples
All three examples we set out below – Dark Nudge, Sludge, and Choice Overload – are seen as likely to be harmful based on the academic literature.
Dark nudge describes instances where friction is removed from a consumer’s journey to allow them to make inadvertent or ill-considered decisions. For this practice, the CMA specifically calls out the gambling sector in its evidence review:
“Another domain where dark nudges are especially prevalent is the gambling industry … electronic machines optimise each step of the gambler’s experience by removing friction from the gambling experience through touchscreen buttons that minimise the physical effort of long gambling sessions …. with the jump to “remote” online and mobile gambling, gamblers today can overcome physical limitations and bring those activities into the home and on the go …. generating a new dimension of gambling harm.”
Sludge describes the practice of placing onerous steps – “excessive” friction – within consumer journeys that makes it difficult for consumers to get the outcomes they want. This is another practice which the CMA speaks about in the context of the gambling sector, referring to its joint investigation with the Gambling Commission into bonuses and promotions online that concluded in 2019:
(Cases so far include) “online gambling, where our concerns included the use of ‘sludge’ and the potential for bonus promotions to be designed in ways that commit people to repeat wagering.”
Choice overload describes the practice of providing so many options to the consumer that they find it difficult to compare, with this difficulty increasing as the number of choice parameters (e.g. price, quality features) increases. The theory of harm here is that overreliance on simplifying heuristics to shortcut this decision making can lead to the consumer making poor choices. In its detailed evidence paper, the CMA uses a stylised example of an online book retailer to demonstrate choice overload (figure 1).
It is hard to look at this stylised example and not draw parallels with online gambling markets, specifically the slots lobby tabs of operators. Interestingly, one of the potential remedies cited by the regulator to mitigate the potential harm is the use of personalisation algorithms:
“Using consumers’ past behavioural data or consumers’ explicit input of preferences, these tools can build short, ordered lists of alternatives that closely match consumers’ preferences. Such tools have been found to improve both the quality and the efficiency of purchase decisions, by enabling consumers to focus their evaluation on the smaller set of high-quality options while lowering search costs.”
This creates somewhat of a dilemma for an industry where the use of personalisation algorithms has come with a degree of controversy. Being able to demonstrate positive outcomes through personalisation is therefore important.
Choice information
Similar to choice structure, the CMA finds that there is strong evidence that manipulating choice information can affect consumer choices. If the information about the available choices is hidden, presented in a misleading way or made difficult to understand, the consumer’s ability to comprehend and evaluate aspects of their choices is weakened and hence they may make poor decisions. Some practices, such as framing and referencing, can be used to affect consumer decisions in a harmful, as well as in a beneficial, way. However, other choice information practices have been found to be harmful most of the time because they may mislead, confuse, disengage or in other ways harm consumers’ choice.
To deal with the harmful choice structure practices, potential remedies typically aim to ensure accurate and unbiased information is provided at key points of consumers’ decision-making processes. However, the remedies should consider other factors, such as information overload or complex language, so as not inadvertently to make the situation worse. In the UK, a form of reference pricing – misleading representation of price or misleading price promotion – may already fall under consumer protection regulation as an unfair commercial practice that misleads or is likely to mislead and hence may lead to enforcement cases.
Table 2: Choice information examples
OCA practice | Description | Evidence |
Drip pricing % | The choice architect initially shows only part of the price and reveals the full price of the product or service at later stages of the consumer journey | **** |
Reference pricing | The choice architect displays a previous (or future) price with the current price, which makes the current price look more attractive | **** |
Framing | The choice architect decides how any decision-related information is described or presented to a consumer | *** |
Complex language % | The choice architect makes information difficult to understand by using obscure word choices and/or sentence structure | *** |
Information overload % | The choice architect gives a consumer too much information about a product or a service such that information about the most relevant attributes is difficult to find and assess | *** |
Choice information – betting and gaming examples
Below, we discuss the use of reference pricing and framing in the sector. Note that these examples are not always seen as harmful in the literature.
Framing is a practice used to describe how operators describe or present decision-related information to the player in a particular way. Behavioural studies show that consumer behaviour can change, potentially materially, depending on how information is framed. At a previous KPMG eSummit, for example, we ran a live framing experiment in the room, finding that even seasoned gaming professionals were not exempt from exhibiting biases in their behaviour when choices were framed in certain ways.
There may be clear positives to framing in the sector. For example, positive framing of responsible gambling tools may lead to a greater take up (and, in turn, consumer protection). However, there have already been cases of framing leading to issues in the sector. The recent UK Advertising Standards Authority investigation into Skill on Net Limited, which focused on the use of a “hot or not” mechanism, included concerns that the way in which information was framed to customers inferred that historical information could be used to assess future performance of games that rely on independent events – preying on the “gambler’s fallacy”.
Reference pricing describes the display of previous (or future) prices alongside current prices. The reference price is seen as creating an “anchor” in the consumers mind and, in turn, the current price is often made to look more attractive.
Markets with transparent price movements have been studied for some time (see the concept of “herding” in behavioural finance, for example). In the gambling sector, especially sports betting, markets look somewhat similar – prices move all the time. However, only recently has the emergence of reference pricing as per the CMA’s definition gained in popularity. So-called “price boosts” and “odds boosts” are now prevalent on many operator’s websites, providing the potential customer with previous prices (the reference price) and boosted odds, often for specific selections as opposed to entire markets. That is not to say that the practice itself is necessarily contributing towards any consumer harm relative to a counterfactual whereby the practice did not exist, but it is worth considering how (if at all) it distorts outcomes.
Choice pressure
The CMA finds that there is evidence that choice pressure can have a strong effect on consumer behaviour. None of these practices are considered to be almost always harmful and each of them can have benefits. However, the CMA also notes that there is generally less existing academic research into them in the context of consumer harm than into choice structure or information practices.
Choice pressure practices may exert influence on consumers, resulting in impulsive, misled and/or unsuitable purchases. In particular, misleading or fake scarcity and popularity claims and fake reviews (a form of messenger practice) can be particularly harmful to consumer choices. Both these practices have been investigated by the CMA. False scarcity claims are already prohibited under consumer protection legislation, while fake reviews are expected to be added to automatically unfair practices under consumer law legislation.
The CMA notes that negative effects of commitment and feedback practices are relatively under-researched and most of the studies explore the potential positive benefits, rather than any potential harm. It is worth noting that, the CMA’s paper cites the gambling sector in its list of examples relating to these two practices – including the use of bonusing in commitment, and losses disguised as wins in terms of feedback mechanisms – both of which could be seen as negative.
Remedies vary across practices but generally aim to ensure consumers are not excessively pressured and misled in their decision making. Remedies also include providing relevant information to consumers and tools enabling them to control the exposure to these practices and their autonomy (e.g. control settings, frequency of reminders, etc.).
Table 3: Choice pressure examples
OCA practice | Description | Evidence |
Drip pricing % | The choice architect initially shows only part of the price and reveals the full price of the product or service at later stages of the consumer journey | **** |
Reference pricing | The choice architect displays a previous (or future) price with the current price, which makes the current price look more attractive | **** |
Framing | The choice architect decides how any decision-related information is described or presented to a consumer | *** |
Complex language % | The choice architect makes information difficult to understand by using obscure word choices and/or sentence structure | *** |
Information overload % | The choice architect gives a consumer too much information about a product or a service such that information about the most relevant attributes is difficult to find and assess | *** |
Choice pressure – betting and gaming examples
The one example we discuss here is scarcity and popularity.Many gambling markets are naturally time limited – the chance to bet the Gold Cup at Cheltenham is limited to once a year, lottery draw closes sales by a given time each draw, and so forth. Historically, it could be argued that online casino games were exempt from this. Recently, however, a new form of casino gambling has emerged – “must be won” slot games. These timed slots have jackpot values that must be won by customers within a given time period (e.g. within an hour, by midnight, etc), creating a new dimension for regulators to consider going forward.
Again, at this stage it is unknown whether these new games are simply innovative concepts in a competitive market that are expanding consumer choice and quality, or whether the use of scarcity leads to demand that is potentially harmful, but it is another example that is perhaps demonstrative as to why the CMA cites gambling so often in its report.
Further consideration – vulnerability
Before moving to conclusions, it is worth noting the CMA’s comments on vulnerability (with vulnerability also being a hot topic at the UK Gambling Commission – for example, see the recent customer interaction guidance). It notes that OCA practices can disproportionately affect vulnerable customers, and that:
“OCA practices used by businesses that rely on repeated engagement, such as gambling or gaming, can be particularly harmful for people at risk of addiction or who are less able to make good decisions, for example, because of age or health.”
What’s next?
The CMA has pledged to investigate harmful OCA practices ‘more actively’, using a full range of powers and tools available. The stronger enforcement powers in consumer protection mean that the CMA could take more consumer protection work and will be able to impose fines on companies breaching the fair use of digital tools if these are deemed to break the law. It will also be undertaking a programme of work that seeks to raise business and consumer awareness of these practices.
More practices, deemed automatically “unfair” in all circumstances, could be restricted or prohibited. Banning or restricting the use of OCA practices could be incorporated into the Digital Markets Unit’s enforceable codes of conduct or pro-competitive interventions. The unit, set up within the CMA to promote competition in digital markets, will be given statutory powers.
References
UK Competition and Markets Authority, Online Choice Architecture, Discussion Paper (April 2022)
UK Competition and Markets Authority, Evidence Review of Online Choice Architecture and Consumer and Competition Harm (April 2022),
Department for Business, Energy & Industrial strategy, Consultation outcome. Reforming competition and consumer policy; government response (Updated 20 April 2022).
Newall, Philip, Dark Nudges in Gambling (2019)
Kahneman, Daniel and Tversky, Amos, Judgement Under Uncertainty: Heuristics and Biases (1974)
Advertising Standards Authority, Ruling on Skill on Net Limited (May 2022)
GB Gambling Commission, Customer interaction: formal guidance for remote gambling operators (June 2022)
Kumar, Satish; Behavioural Biases in Investment Decision Making – a systematic literature review (2015)
The Consumer Protection from Unfair Trading Regulations (Statutory Instrument No. 1277, 2008)