Editor's note: Jon Puleston is vice president in the London office of Seattle research firm Global Market Insite Inc. He can be reached at jpuleston@gmi-mr.com. This article appeared in the January 23, 2012, edition of Quirk's e-newsletter.
Gamification is one of the buzzwords in the marketing industry right now and is a subject of growing interest amongst market researchers. The idea is being applied to everything from teaching to brand marketing, even to whole social engineering schemes by governments. Over the last year, GMI has conducted extensive research looking at how to apply gamification techniques to make surveys more fun and measuring the impact they have.
In this first installment of a two-part article, I will show you how the thinking and theory behind game play can be applied in market research, demonstrating the dramatic impact that gamification can have in improving the survey experience for respondents - and the quality of feedback. The second part will focus on designing questions to be more game-like.
To begin, gamification is the process of applying gaming mechanics to everyday tasks to encourage more active participation. The theory of gamification: more fun = more feedback. There is a vast amount of literature on the theory of game play but if you are interested in pursuing this further I suggest, as a starting point, that you read Reality Is Broken: Why Games Make Us Better and How They Can Change the World by Jane McGonigal. Or, simply type in gamification on Google.
What defines a game?
The starting point for all this is to understand what turns an activity into a game. There are all types of activities that we could call games, some more obvious than others. A game is just about any form of thinking activity that we do for fun and that could encompass playing conventional games such as I spy, chess, cards, sports or video games. But in pure game theory terms, watching TV, gardening, cooking, photography and even sex could all be seen as games.
There is often very little to differentiate work tasks from games other than an element of fun and a few silly rules. If I asked you to carry a 10-pound bag for five miles you would probably expect payment. However if I give you a stick and a challenge to hit a ball with it along the route in the minimum number of touches, it has been transformed into a game called golf, which people pay to do.
Think of the hours kids spend playing computer games that often center around quite serious problem-solving, requiring the same intellectual input as doing math homework would demand. Or take our willingness to climb up mountains for the sake of challenging ourselves in comparison to our unwillingness to walk any sort of distance if you feel you are being forced to do it, at an airport terminal as an example. You could perceive of a market research survey as being a game too though unfortunately, a rather boring one in most cases.
The impact of making surveys fun
The main problem we face with online research is motivating people to participate in surveys while giving us their full attention. The key reason for this problem is simply that surveys are generally seen as boring. In a recent study published by Survey Nation, 58 percent of respondents said they don't like doing surveys.
We play games for fun, for entertainment, so the first hurdle to cross when repositioning surveys into more game-like experiences for the respondents is to ensure that they come across as a fun process. We have discovered that by telling respondents from the outset of a survey that we would like them to play a survey game - as opposed to completing a survey - that there is a measurable transformation in the respondent's attitude and approach.
Here is an example: We conducted an experiment whereby we asked two groups of respondents to take part in an identical media touchpoint survey where we asked them to evaluate their consumption and use of a range of different media. In one cell we presented the survey in the usual traditional way and in the other we told them they were about to play a game where the objective was to plan an advertising campaign.
To reenforce this, we created a customized introduction page and made a few adjustments to the wording of the questions to make them sound more game-likeure 1). However, the tasks they were asked to perform were essentially identical.
We discovered that during the game-style survey, respondents spent 20 percent more time answering like-for-like questions, with their overall rating of the survey experience at the end increasing from 61 percent to 84 percent saying they enjoyed the survey.
Rethinking question design
The problem of surveys being seen as boring centers around the questions themselves and they way they are asked. So often the questions we ask in surveys are dry and emotionally unengaging. To make surveys more game-like you have to rethink the approach to writing questions and the style of questions.
The heritage for the language used in surveys dates back to the era of human-based (face-to-face) interviewing techniques where the need to engage respondents was hardly an issue. The more important factor was being clear and understandable and, without any visual clues, the language of surveys also had to be very descriptive. This led to the adoption of verbose, overblown phrasing of questions. This can make reading survey questions seem more like reading a legal document - an emotionally alienating experience. It is crucial that a more succinct, fun, human and engaging approach is employed in online research.
With human-based interviews you don't have the same worry that online research has as to whether the respondent is actually listening to the questions or is bothering to answer them properly. You are there with them; you can see if they are paying attention. Rarely during human-based interviews would you expect respondents to walk away halfway through an interview without saying anything at all to any questions or repeatedly giving the same answer. But with a computer it is all too easy for respondents to give only a fleeting, momentary thought to the questions, make up any answers, tick any option and click on to the "Next" button or simply close the survey if they get bored.
There are many different methods based on human psychology and theory to increase a survey's entertainment value, including changing the question style, applying abstract rules to questions, sending respondents on quests, scenario-planning, competition, rewards and involvement.
Changing the question style
If you are thinking about gamifying your survey, to begin with you need to think about adopting a more engaging approach to the way you word the questions. Put yourself in the mind-set of a good TV interviewer like Oprah: Would she start an interview by asking, "On a scale of 1 to 10?"
There are various techniques that can be used to encourage respondents to want to read and answer questions:
- Personalization
- Emotionalization
- Projection
- Forced imaginary situations
- Fantasy
Personalization. One of the most powerful techniques you can use to reframe questions, to make respondents more interested in answering them, is to make the question seem like it is all about them.
A question such as, "Which of these is your favorite color?" could be personalized by asking, "If you had to paint your room in one of these colors, which one would you pick?" Now, I recognize that the question has changed and will affect the overall character of the answer but how you personalize questions can be very easily adapted to the relevance of the questions.
We have found that respondents will spent upwards of 20 percent more time thinking about questions when they are reframed to include a more personal context.
Emotionalization. (Appreciate this is a made-up word!) This is about triggering some latent feelings that would encourage respondents to truly think about the question.
An example of this might be changing a question that reads, "What would you wear?" to "What would you wear on a first date?"
This, too, is a form of personalization but one that targets emotion. We have found that an emotional trigger in a question can encourage up to 50 percent more consideration time and feedback from respondents.
Projection. This is a common technique used in qualitative research that involves asking the respondent to imagine something in the mind's eye of someone else. This is a very powerful engagement technique to use in online surveys. For example, one question might be, "What do you think about this new product?" Applying projection to this question, it might read, "Imagine you are the boss of a company. Your job is now to evaluate this new product..."
We have conducted various experiments using this technique and found that it can easily encourage up to twice as much attention and feedback from respondents when used in the right way.
Forced imaginary situations. Forcing imaginary situations on respondents when done in a creative way can really help respondents stop and think about a topic. In relation to a clothes-choice question, you might say something along the lines of, "Imagine you spilled coffee all over your pants and had to quickly go out and buy a replacement pair..." The idea is to add a slight twist to differentiate the question and make it stand out.
Use of out-and-out fantasy. The next level up from this is the use of out-and-out fantasy. Take, for example, the earlier question, "What clothes would you wear?" To include fantasy, ask: "What would you wear if you were going to appear on TV?" "What would you wear if you had all the money in the world?" "Imagine you won a trip to Paris and part of the prize was a clothes shopping spree..." What we have found in our research is that the more imaginary the framework of the question, the more respondents enjoy answering them and the more feedback they will give you.
Applying abstract rules to questions
Abstract rules are a core constituent of most games. Rules can be used to transform almost any task into a game. The trick to doing it effectively is to make them strict (more on that later), silly, abstract and irrational. If you scientifically analyze what differentiates games from work it all boils down to rules and rewards.
Some of the techniques outlined above to make questions more engaging are rule-setting processes, if you think about it in strict theoretical terms. We have found that applying strict rules to how respondents can answer questions can transform questions into games in the minds of respondents.
Here is what is now quite a famous example:
Question: Please describe your favorite meal.
Question with rule: "Imagine if you were on Death Row and had to plan your last meal. What would you order?"
As the examples in the accompanying chart show, the question without the rule of planning a last meal yielded an average of three words per respondent, compared to 15 words per respondent with the rule included.
There are many different ways rules can be applied to questions. Here are some techniques we have discovered work well.
Boiling down to specifics. Adding a specific scenario to a question is the easiest and most versatile rule technique you can use. Many people do this naturally. It is more fun to answer if you answer the question in a specific context: What would you wear last year that you would not wear this year? What would you wear if it were a hot, sunny day? What would you wear if you were going to a job interview?
Restrictive rules. This a technique where you place a limitation on what respondents can do. As an example, you can restrict the number of words they can write, which can have quite the opposite effect to what you might expect. If you ask respondents to describe themselves, on average they write 2.4 descriptors. However if you were to say, "Describe yourself using only seven words," which reads like a restriction, this actually results in an average of 4.5 descriptors.
Whittling down rules. Forcing respondents to make decisions is another interesting way of asking people to respond to questions. For example, one question might be, "Out of your clothing collection, if you had to pick five items of clothing that reflect you best, which would you pick?" Or, "If you were to go on vacation and you could only take three outfits with you, one for daytime, one for evening wear and one for a special occasion..."
Sending respondents on quests
If you study the evolutionary roots of game play it is believed that our propensity to play games evolved to help us hone and develop our survival skills, particularly our hunter-gatherer skills. Hunter-gatherer missions could last several days, during which time we learned to maintain extended periods of concentration and dial down all other distractions on our brains. We used games to practice this.
Many of the best games flip people directly into this hunter-gatherer mind-set. Think about some of the most successful video games such as Call of Duty or World of Warcraft where they encourage people to spend hours at a time often doing quite mundane tasks in an effort to complete a quest mission. We discovered that this same thinking could be applied directly to question design. By rewording questions to come across more like quests and missions, we found we could significantly increase the amount of time respondents were prepared to spend answering them.
A question such as, "How much do you like these music artists?" could be reworded to, "Imagine you owned your own radio station and could play any music you liked, which of these artists would you place on your playlist?" The number of artists respondents were prepared to rate increased from 84 to 148. The wording had made it more relevant to respondents. It had a point. It had become a quest and as a result they focused more on the task and spent twice as much time on it.
This is a very versatile technique that can be applied in all manner of different question circumstances. It is essentially about adding a reason for completing a task that means something to the respondents.
Scenario-planning
In a similar vein, the frontal cortex of our brains has specifically been developed to plan scenarios. Many games are centered around scenario-planning, what-ifs and other imaginary situations. Changing a question to evoke a scenario-planning process is another way of making questions more game-like. Instead of asking what words a respondent would use to describe this brand, you could ask, "Imagine this brand was a human being, what words would you use to describe this person?" We have found that this twist in wording encourages respondents to choose up to twice as many characteristics.
Working with Kimberly-Clark, we built a survey around scenario-planning as a means of segmenting attitudes toward a pair of competing brands. We asked respondents to take part in a series of mini games where they had to guess which brand was most likely to have different human characteristics and they were to create personality portraits of each brand using visual imagery. This can be a notoriously difficult task for respondents to process, but couched in the framework of a game they are better able to conceptualize the task. It helped to pull apart differences between the brands in a way that is hard to achieve using traditional questioning approaches.
Respondents found this way of thinking about brands enjoyable: 97 percent who started the survey completed it; 78 percent of respondents rated the survey four stars or higher; 90 percent said they felt they were fully engaged in the survey; and 93 percent said they would like to do more surveys like it.
Adding a competitive element
Most games invariably involve some sort of competitive element. We have found adding any form of competitive element to a survey seems to provoke a strong positive reaction. For example, instead of asking how many brands of insurance companies a respondent can remember, you could add a restrictive rule such as a time constraint. If you tell respondents they only have two minutes to answer the question, respondents think this does not give them much time and they rush to give you answers. In short, they treat it as a challenge.
The reality of this is that a two-minute time limit is not a constraint at all. Normally respondents might typically spend only 30 seconds answering a question like this. In one experiment where we asked respondents to name foods they like to eat, we moved from six foods listed to 35 when we imposed the time constraint.
Offering rewards
The reason why we play most games is for the reward. This may be simply the positive feeling you have for getting an answer correct or from winning something tangible like points or prizes. Adding reward mechanics into surveys is one of the most powerful means of gamifying a survey experience. There are several ways this can be done. The areas we have explored have been giving people points for making correct predictions and trading-style games where respondents are asked to gamble on their own opinions. Both of these techniques can evoke twice as much time being spent on a task and significant improvements in survey enjoyment.
As an example, we conducted an experiment where respondents had to predict the future of different brands. To make this process more game-like we gave each respondent a trading budget and asked them gamble on their opinions. If the "market" agreed or disagreed with their opinions they won/lost the money they had gambled.
We found that in this mindset respondents easily spent 100 percent more time thinking about what answers to give. The essential balance of the data was very similar but the data from the gambling mechanic also added a new level of insight, which was a measure of respondents' confidence in their opinions. When asked about the experience of answering survey questions in this way the feedback was almost universally positive. Over 90 percent said they enjoyed it.
Making tasks more involving
Nearly all these game play mechanics actually make the process of answering questions slightly harder. But if you think about the success of some of the most popular games, nearly all of them call for a lot more thought and effort than we expect of people when they complete surveys (e.g., Scrabble, chess, most card games, etc.).
So often when we design surveys we are fixated on asking respondents simple questions rather than challenging them to do more complex thinking. We have found that the more complex a task we give, the more fun, enjoyable and rich an experience this can make the survey for them.
Working with Sony to develop an ethnographic research study, we added a rule that respondents had to imagine they were being interviewed by a magazine and that what they wrote would be published. The impact of this was that respondents spent 30 percent more time answering the survey and overall word count increased from 230 to 350 words. It was a far more interesting task when reframed in this way.
The sweet spot
For games to be successful they need to be viewed as a possible challenge with the right balance of luck and skill. Make a game appear too difficult and few people will want to do it. On the other hand, make a game seem too easy and again, few people will be motivated to play it. So when you design question games, you need to find the sweet spot between easy and difficult.