The Problem: How it All Started

The hobby of board gaming is currently experiencing a renascence. check out BoardGameGeek here BoardGameGeek, the biggest board games database, has more than 70 000 games. Out of those, hundreds are worth playing. I got into this hobby four years ago and I then hit a wall. When I went to play games together with my friends we would spend hours choosing what to play. I felt I was missing out on good games and time I could spend playing.

I first had the idea of Tabletop Match when I saw an enormous decision tree meme that was meant to help you 'choose the perfect board game'. "This can surely be automatised" I told myself. So when I had to develop a mobile web app with strategical content for my Bachelor thesis I figured: "This is is!".

Understanding the Users

Inspired by Cooper’s book, About Face, one of the best written design books About Face, I wanted to apply Goal Oriented Design practices. So, I went to find out who my users are and what they need. It was the first time I conducted exploratory interviews, so my questions were pretty shaky:

• survey-like: How often do you play board games?,
• yes/no questions: Do you have any favourite game?,
• leading questions: Would you find useful answering a questionnaire before playing, which would give you as an answer a set of possible games you could play?
• or even unclear subjective questions: Would you consider yourself an expert or a casual player?
• but also very useful questions: What do you do when you want to try a new game?

To find interviewees, I went to a Disambiguation: 'Tabletop game' is a general term that refers to board games, card games, dice games, miniatures wargames, tile-based games and other games that are normally played on a table or a flat surface tabletop club and a hobby store in Herning, Denmark. In total I interviewed 15 men and one woman. Most of them were members of tabletop gaming clubs.

After a few interviews I quickly figured out some mistakes in my questioning, and I reformulated. But I did not fully understand how to devise good interview questions yet. Nonetheless, the data was interesting and patterns and personas evolved from it.

Fig 1. I first identified significant behavioural patterns, then I mapped the interviewees accordingly. In the end I indentified three groups which became the basis for personas

From this, I noticed several correlations between behaviours:

• Players who don’t belong to a club tend to know fewer games than players who do.
• Players who know a lot of games prefer to actively search for them. They are expert gamers and tabletop games are generally their main hobby.
• Players who belong to clubs tend to rely more on recommendations from others when trying new games

This helped distinguish three personas. A fourth persona was developed with the help of a Read more about Dan Shapiro and his board gamers venture here. New York Times article about how board games are very popular among parents and children who want alternative entertainment to computers.

Starting up anew

One year after the first iteration, I took up the challenge of Tabletop Match again. This time as part of my Master’s degree, together with a team of three, we had to build a viable tech business & product. My colleague, Andres Sierra, had always wanted to build a start-up around board games, as he’s an avid player and an entrepreneur. So we teamed up with Francesco Vitale, the amazing designer, and Wilfried Bock, the team's hacker, to bring Tabletop Match to the next level.

Arising Questions

• Does the initial concept of Tabletop Match really answer the needs of gamers?
• Are gamers our target market, or maybe someone else (game makers, game shops, game publishers)?
• How can we deliver value to users and sustain a start-up?
• Is this market big enough to build a business upon?
• How can we make profit?
• How can we grow?

A Business Approach

This time we used a business oriented process: Read more about the Lean Start-up Method the lean start-up methodology . We started by formulating and testing our assumptions using the Read more about how the Business Model Canvas is used to build great companies Business Model Canvas .

Along the process we changed the BMC 7 times until we reached a viable business model

We then devised tests to validate each of our assumptions. For that we interviewed 9 board gamers and 3 owners of hobby stores.

Our assumption was that hobby stores would like to pay Tabletop Match to feature them as places where users can buy the games they want. But the stores we interviewed did not seem to need or want to pay for online advertisements.

But what we found is that with all the new games on the market, it was nearly impossible for hobby store employees to keep up. Yet customers heavily relied on them for advice. Thus shop assistants liked the idea of having a filtering app in their store to help them browse, understand and recommend games.

Customer Archetypes Diagram: Based on interviews, we divided our customers on two behavioral axes. This division is somehow similar to the personas developed earlier.

We decided to focus on casual gamers as our main user group and on board game stores as our customers. Tabletop Match pivoted to be a game inventory management system for hobby stores.

On top of that, the system has a customer-facing app on a tablet installed in the store. On this app, hobby store customers can filter through games and find the right match for them.

Customer Workflow inside a hobby store

A New Web App

We built a new web app to showcase our customer-facing interface concept. This time, Tabletop Match was built to be used on a tablet or desktop primarily.

The problem with the initial matching method, multiple choice questions, was that the user had according to Miller's law, one can pay attention to seven ± two items at once too many choices (7 to 11).

To solve that, and inspired by reviewing one choice at a time helps users focus, and is a popular design pattern Tinder , in the new system one can answer a series of yes/no questions together with some hard filters to get a match. But this soon turned out to be a terrible design decision: a user would have to go through too many yes/no questions to get a relevant recommendation. In hindsight, a better solution would be to use multiple choice questions, like in the first iteration, but limit the choices to according to a more recent study, the number of items one can actively hold in working memory is actually four four .

We eventually figured how to solve one of our biggest challenges: building a proper database. By using each store's own database, we would not have crowd-source it or build it from scratch. Once we have access to several databases, we can cross-check them and complete the missing details.

Check out our company's promotional video

Killing our Darlings

The end of the university year arrived, and as a team we had to part before the project was complete. We left a lot of loose ends:

• designing a new multiple choice system
• releasing a beta to the public
• building the inventory management system
• building ways to migrate databases
• finding more customers (hobby stores)

As a team, we could not take Tabletop Match further, as each of us would move to another place. Moreover, we did not have the necessary technical skills to build database migrating system. Thus, Tabletop Match became again a personal project, for us to take further in our free time.

Lessons Learned

Tabletop Match is the project that I have grown with and has taught me the most about design and entrepreneurship. It was my sandbox, where I experimented and put into practice design theories and methods I learned from books and in university.

Through Tabletop Match I discovered my passion for understanding the users and their needs, for learning their stories and translating that into actionable design solutions.

Even if it was never released or launched as a start-up, the team has worked for Tabletop Match in the true spirit of a lean start-up. We built hypothesis, invalidated them, pivoted and took it dead serious until the very end.

In the end, I also learnt the hard lesson of loose ends and killing your darlings. It was difficult to let go of the project that I have grown with.

Curiosity Killed the Cat

During the past two project iterations we conducted 25 interviews with board gamers from Denmark, Sweden and Romania. But back then, I was still learning about data analysis. Moreover, I’ve never really put all the data in one place.

So, one year later , when I was writing this case study, the curiosity got the best of me. This time I knew so much more about how to extract meaningful results from interviews, after working in an read more about ExSitu and its projects here HCI research lab for 6 months. So I wanted to go back to my data to find out how do board gamers really choose which games to play and Discovering New Games. I wanted to understand if Tabletop Match would be an useful app for the community and whether it is worth continuing to work on it.

Methodology

Fortunately, in both interview sessions we asked questions about:

• How often do they play games
• How many games they’ve played
• Playmates
• Discovering New Games
• How do they chose a game to play

... and this made it possible to put the two interview batches together and treat it as one body of data.

I went back to the interview transcripts and put them into a better format for analysis. To understand the level of involvement in the hobby for each participant I used the number of times they play per month and the approximate number of games they've ever played.

You can browse the data here

The rest of the data was qualitative, so I used the imported from social sciences, grounded theory is useful for analysing qualitative data from user interviews. Read more about how it works here. grounded theory method to analyse it. I first chose each relevant quote from the transcripts and I labelled it with a code. Several concepts arose from the data, which I then divided into four categories:

What I play
Who I play with
How I discover
How I chose

You can browse the data here

Within these big categories I found several dimensions. For example, in the ‘What I play’ category, respondents:

• either had few preferred game they always played, and weren't interested in discovering other games,
• either had one preferred game but were open to explore more games,
• or their main enjoyment of the hobby stemmed from finding new games.

Results

By combining several behavioural variables I finally grouped the 25 interviewees into 4 user categories: