Can gamification datafy human thought, as well as human behaviour?
Datafying human behaviour is the next frontier for analysing unstructured data and prospering in the knowledge economy. But the final frontier is using digital techniques such as gamification to datafy human thought - the memories and perceptions; ideas and ideals; attitudes and opinions; preferences and values that form, morph and re-morph in our cerebral cortices and underlie every decision that we make.
The value of this information has long been recognised and traditionally been sought using questionnaires. Customer satisfaction surveys are a prime example, the irony being that these surveys often enhance dissatisfaction through providing a poor customer experience. By failing to put themselves in customers’ shoes and only focusing on what they are after, researchers create an interaction that provides customers with no pay-off for giving up their time. Not surprising then that response rates are poor and data quality is questionable.
Gamifying customer research to improve response rates and quality
Using extrinsic motivation such as paying customers a fee or making a contribution to charity is one way to increase response rates, but extrinsic motivation doesn’t necessarily improve quality, arguably it can have a negative effect. Respecting customers’ time and ensuring they see a clear link between providing answers and receiving better service is the ideal response. But some important questions that managers need answers for only benefit the business, for example those relating to brand awareness and the allocation of advertising expenditures. With these questions the only option is to increase engagement and one alternative that is particularly suited to online surveys is gamification.
Gamification uses gaming psychology to engage respondents and encourage the desired behaviours. It taps into basic motivations such as the need to control, learn, achieve, contribute and interact through the creation of a clearly defined goal, with progress towards it being measurable and feedback via rewards or recognition for each goal achieved, both intermediate and ultimate. Gamification has typically been used to encourage customers to self-serve or help each other; to motivate sales staff to complete training, enter data into the CRM system and share knowledge; and to encourage all staff to collaborate through contributing to community sites. But increasingly its possibilities in capturing data are being recognised.
There are a number of principles underlying successful gamification. Firstly, the focus needs to be on learning which requires a challenge that matches a participant’s ability and increases as ability improves. Secondly this challenge is coated in an engaging story – a theme designed to lure participants into the game. The conditions for success – progress through levels – have to be clearly defined with a feedback loop to communicate relative and absolute progress (via a leader board and badges respectively), both of which can be shared so status can be celebrated. (Badges can also be used to reward desired behaviours even if they do not result in measurable progress.)
Progress also should be non-linear, incorporating alternative story lines and mini games - side tracks that the participant can choose to follow or not – to keep the experience fresh. Non-linearity also introduces uncertainty so guesswork based on repetition of previously successful strategies won’t work, ensuring that learning occurs at each level. Also, the temptation to cheat will be difficult for some people to resist, so providing scope for doing so in a way that is compatible with the learning objectives is sensible. Finally, there will be unintended consequences – people behaving in unexpected ways because incentives have had an unforeseen impact – so measuring, iterating and improving on a regular basis is paramount.
These principles can be applied to online research to improve response rates and the quality of responses. This applies whether the insights are to be gained via user stories posted on a community forum – with badges for best story or best summary of all the posts on a particular topic – or through an online questionnaire, where low rates of response and speeding or selecting answers from the same column are all too common.
When there is no resulting pay-off to answering questions, making the experience engaging is the only way to reward participation. At the most basic level, this can be achieved by adjusting how questions are asked. For example, an element of challenge can be introduced by setting respondents a time limit for completing the response (name as many beers as you can in two minutes) or word limit – in no more than seven words describe how you feel about the service you have received. Then there is specifying a particular context – asking what clothes someone would wear on a first date to elicit what fashion brands make them feel good rather than a vaguer question about brand preferences. Introducing an imaginary quest – being asked to imagine that you are writing a recommendation that will be published in a food magazine or creating a playlist for a radio station –also makes collating and expressing thoughts more fun than simply being asked what restaurants you like or what music you like to listen to.
Online surveys also lend themselves to more visual and immersive selection processes – moving an image from one side of the screen to the other or placing a cross or tick over it. Likewise, they enable autonomy and non-linearity through the incorporation of routing questions which allow participants to choose the order in which they answer sections. Finally, online surveys where there are right or wrong questions –matching tag lines to brands – can be gamified by introducing immediate feedback as to whether the answer was correct or through the award of points for every correct answer.
Prediction markets – gamifying staff insight collection to improve strategic decision-making
Anglo-Saxon business culture is moving away from the view that strategic decision-making is the preserve of experienced executives supported by a small cadre of advisers. IBM pioneered the use of company-wide ‘Jams’ – centrally managed bulletin boards seeking substantive answers to business-critical questions in a defined time period (e.g. three days) – to generate new ideas and engender participation. Jams leverage the wisdom of crowds – the idea that collectively we are smarter than the smartest individual so long as the behaviour of one participant does not influence that of another.
An alternative, gamified way to collect employees’ insights and intuitions is via decision markets (also known as prediction markets). Decision markets generate crowd-sourced predictions of future outcomes by aggregating opinions using a stock market mechanism. Market participants buy or sell shares in ‘claims’ about a particular prediction. Popular examples include the outcome of elections, Oscar winners and the box office takings of new films.
The same approach can be used to create broader-based and more informed predictions of business-critical events. Employees can buy either ‘yes’ or ‘no’ shares, depending on whether they believe an event will transpire or not, for example whether a new product launch will generate a threshold level of sales (e.g. $50m) in its first year. They can make money by either holding on to the shares until the claim has proved to be true and receive the promised pay-out or by selling the share in the market if they feel it has become overpriced relative to the likelihood of the claim’s success.
For such markets to work, the uncertainty needs to be distilled to a simple choice – for example which of two or three technologies will prevail – and for the results to be measurable, e.g. the technology becoming declared the industry standard by a qualified body. (Where such definition or measurability is less clear, it can be manufactured through the incorporation of timescales and the creation of a threshold.)
Aggregating opinions produces a forecast in which more confidence can be placed. There is a long history of research showing that the average of multiple predictions generally outperforms individual predictions; and more recently that crowdsourced predictions generally outperform those of individuals, no matter how smart the individual is. The profit mechanism ensures that people with good insight are rewarded and have more funds to invest in new claims while those with only poor predictive abilities are limited in future trading by lack of resources. Most important of all, there is no hierarchy – the purchase or sale of a claim by a senior manager does not have a disproportionate effect on price movements because of their seniority.
The increasing volumes of data that is available should not blind us to the fact that what is available is not always what is required. There is a valuable trove of information in the heads of customers and employees which, if extracted, can significantly improve the quality of decision making. Traditional research techniques can work, but typically suffer from both poor response rates and poor quality responses. Gamification can change that and works well in digital environments. It doesn’t overcome the manufactured situation in which questions are being asked, but play-like participation helps distracts the conscious mind from the sense of confection, freeing the sub-conscious to communicate what it really thinks and feels.
Jack Springman is head of customer analytics at Sopra Steria.