How to design a game to teach creativity
Why creativity matters
Software is eating jobs. Even high-skilled jobs that were previously thought of as part of the ‘knowledge economy’ are becoming automated and outsourced. Creativity is the key distinguishing factor between high-paying, differentiated work and work that can be commoditized. It is no surprise that CEOs rank creativity as the most crucial skill for the 21st century workplace. Meanwhile, because schools are killing creativity, many well-educated workers feel unprepared to take on the challenges of their jobs in a fast-changing world.
Who we are
Lemonade.io is a group of educators and technologists united by a belief that it is not only possible but also vital to teach creativity. Our founding team includes David Kelley (founder of IDEO & Stanford d.school and NYT-bestselling author of Creative confidence) as well as Stanford Education, MBA and Computer Science alums. In the past our team has helped build products like Learnist, Reading Raven and Fitbit. We are also supported by influential Silicon Valley investors including NEA and Jerry Yang.
How to design a game to teach creativity
We recently released our first app, Funder on the App Store. Funder is not just a creativity game, but also a hilarious parody of the Silicon Valley startup scene. We had set about designing a mobile game to teach creativity and came up with a notebook full of ideas. After all that iteration, here are our main lessons for designing a game to teach creativity:
1. Use prompts to get the creative juices flowing.
The most obvious way to get people to be creative is to just get them started. We could ask our users to use their phone to draw a picture, write a haiku, record a song or take an interesting picture.
However, we quickly learned about the blank slate problem. We found that people feel stuck when they are given a medium and asked to be creative. Try it yourself: Ask someone to write a poem. Their immediate response will be, “About what?” Counterintuitively, people are much better at generating responses for a specific problem. A smaller set of possible solutions is surprisingly more motivating for people to get their creative juices flowing.
Thus a core element of a creativity game would have to be a prompt that one can respond to. This is similar to writing a clue for a crossword puzzle, but asking people to generate something creative instead of a predetermined answer.
In Funder, the prompt is the name of a real company, and players write fake pitches for it. For example, here’s what a real user came up with for Instagram:
That brings us to the most important question, how do we decide how someone did?
2. Find a way to score creations.
One of the most important elements of a game is feedback. Once I complete a task, how do I know how I did? For a computer to give you a score, it has to know how to measure your performance, but computers don’t yet know how to measure creativity. We believe this is the primary reason why there are no mobile games that center on creativity: most games have a handful of ways to “win” the game, but the only time you need to be creative is when the solution is non-obvious.
However, contrary to what most people think, researchers do know how to measure creativity. The gold standard for creativity assessment, known as the Consensual Assessment Technique, was developed by Teresa Amabile at Stanford and Harvard three decades ago. It basically states that when experts are asked to evaluate how creative something is, their responses are internally consistent and looking at their aggregate choice is a pretty reliable measure of creativity.
We knew we couldn’t get enough experts to keep up with an active online community of creators, but what if the creators could also be the judges? We found that it is much easier for people to identify creativity than to be creative themselves. We began to imagine a social game, where we crowdsource community votes on submissions to decide the winner. We found through user testing that in a game, players expect a score as soon as they complete the task. However, since we needed to give the community time to score creations, we decided to set the expectation appropriately by designing the game to look more like a social network.
In Funder, we used the Silicon Valley mechanic of “investing” in a pitch, which increases its “valuation”. This is a reward not only for the person who created the pitch, but also the people who invested in it early. Users can invest in pitches using the familiar upvote or downvote. In many of the other games we prototyped, we found that it wasn’t nearly as exciting to read content that other people wrote after writing content yourself, but with Funder we found a really good balance.
How do we ensure that those assessing are familiar with the context of the creation?
3. Focus on shared context.
Have you ever tried to explain an inside joke to someone? Then you know what I mean when I say that a big part of creativity is context. Creating something that a specific audience similar to you will like is much more enjoyable than something that the global population will like. Similar to the first lesson, having the constraint of a more limited set of possible things to play with is actually more empowering than having all the tools in the world.
In Funder, people make fun of real companies by writing fake pitches. We decided to focus on the startup scene because it’s a group we know well and one that’s quick to adopt new products. We filled Funder with inside jokes about Unicorns, Self-Driving Teslas and Changing the World to make the experience delightful for them. However, we’ve been surprised to find that the appeal of the app has been fairly universal, with people writing pitches not just about Facebook, Apple and Tinder but also about Donald Trump, TED and Starbucks.
4. Give people reasons to come back and play more.
The best games are the ones that give an incentive to keep coming back. Luckily, a social game based on creativity has a unique set of advantages. Getting people to create something gives them a very special sense of ownership over it. This not only makes them more likely to share it with their friends outside of the game, but also gives them an intrinsic motivation to come back and “check on” their creations. The social nature of the game means that when a user invests in something, the creator gets a notification and a reason to come back. Moreover, even when users aren’t feeling creative themselves, they can come to the game to read and invest in submissions by other users for some laughs and inspiration.
In Funder, we used the analogy of “net worth” to incentivize users to come back. Every user starts off with $1M of fake startup money, and they can grow their net worth as they create or invest in successful pitches. Of course, users can compete with their friends on a leaderboard to get a higher net worth. Moreover, as their net worth grows, users level up in the game. They climb the ladder from Intern to Growth Hacker, Keynote Speaker, VC and beyond. They can show off their level with a cool badge on their profile and in their friends’ feeds. Each level comes with a hilarious description that users can only see when they unlock the level.
Why design a game to teach creativity
Now the most important question, why build a game to teach creativity at all? Even if creativity is a valuable skill, why not teach it with a course or a book instead of building a mobile social game?
Because we have a secret. At Lemonade.io, we set out on a mission to make the world more creative but we quickly realized that it is hard to teach creativity without knowing how to measure it. We also concluded that part of the reason that creativity wasn’t taken seriously was because it was hard to measure. Thus, we decided to focus on building technology measuring creativity and reviewed the research from the psychology side as well as the computer science side. We found that machine learning can be used to teach a computer to apply the Consensual Assessment Technique at scale, but the biggest barrier was the lack of data. Machine Learning only works on very large scale datasets and even the best algorithms aren’t very accurate without enough training data. No one in the world had a dataset remotely large enough to get high accuracy, so we chose to build our own.
We decided to use a paradigm known as Games With A Purpose (explained brilliantly in this talk by its inventor Luis von Ahn). The key principle is that we can solve complex algorithmic problems by breaking them down into tasks that are fun for humans to do. For example, a computer does not know how to write something creative given a prompt or how to assign a score to a given creation. If we can turn that same task into something that people like to do, then we can get them to do it a lot without having to pay them to do it. This is the key principle behind projects like Duolingo, Recaptcha and FoldIt. Most importantly, since people are generating value just by playing the game, we can give it to them for free. I’ve always believed that education is the most powerful vehicle for reducing inequality, and so the prospect of a product that is free for users was really exciting.
In Funder, users see a long list of potential companies they can write parody pitches for. After writing a pitch for a particular company, they can read and invest in pitches that other people wrote for the same company while their own pitch also gets rated by other users. The most exciting thing is that we aren’t just collecting finished-product pitches and votes on those. Since people are creating and voting inside an app, we are able to collect an unprecedented amount of data about the creative process. For example, we record the order in which people write the words of the pitch and all the changes that people make to it before they submit it.
Imagine if Shakespeare did his writing on Google Docs, and we could see his “Revision History”. We’d learn the order in which he wrote his plays, where he made edits or additions, and get to read all the things he decided to edit out. I argue that we would learn more about his creative process than the entire history of scholars have been able to.
(White House photo. See also: “Barack Obama, Editor”)
Response thus far
We launched Funder last month to a small set of users and they love it so far. 80% of our new users come back the same day. The average user who has written at least one pitch has already gone on to write more than 7 pitches. We have thousands of pitches and people have voted on them over 20,000 times — and we’re just getting started.
Above all, our users have surprised us with their wit, creativity and poignancy. You can read and vote on their pitches by downloading Funder on the iOS App Store or visiting our website at funder.lol/. Here are just a few of our favorites:
Lastly, here are some surprising ways people have used the app:
Read and vote on their pitches by downloading Funder on the iOS App Store or visiting our website at funder.lol/
