Using data and empathy to uncover user needs at PaintBerri


PaintBerri is an art role-playing and storytelling community co-founded by Katherine Tung and I in January 2016. The site began as a side project led by Katherine in January 2014, who was inspired by the vibrant art communities on Japanese Oekaki boards (visual forums where posts and comments are drawn with a simple painting program).

With a small team, we have grown the user base to over 23,000 artists which include over 120 paying customers, all without outside funding or paid advertising.

As of August 2016, development on the site is put on hold. Read about the team's journey here on Medium.



We thought artists were coming to PaintBerri to draw with the unique browser painting program, but something didn't seem right

PaintBerri was based off of Oekaki sites, which were popular in America in the late 2000’s. They were a gathering spot for artists to draw online with a simple painting tool, improve their skills, and share art.

Artists also frequently create visual stories on Oekaki sites, called art role-plays (roleplay, RPs). A role-play is an improvised collaborative comic created by one or more artists telling a story through characters they usually design themselves. 

When PaintBerri went into open beta, people came flooding in from Tumblr and one of the last English-language Oekaki sites, Tegaki-E, and we assumed it was because PaintBerri had a more robust painting program.

At the same time, we were drowning in complaints from people who had compatibility issues with the painting program or wanted more features.

Excited by the rush of users, the team decided to respond to popular opinion and churned out a series of painter upgrades and features.

However, despite sinking hundreds of hours into constant upgrades to the painting program and surpassing the feature sets of other Oekaki sites, PaintBerri’s growth was stagnating.


Surveys, interviews, analytics, and a fiery controversy help the team understand what artists really want

Unsure of what was going on, we designed a satisfaction survey and presented it to every logged in user for a few days.

After collecting over 1,300 responses, we dug through the numbers discovered a few interesting things: people who liked the site enjoyed it because of the art role-playing culture, while those who didn't like the site had trouble with the painting program.

If half of the happiest users were happy because they could art role-play and share their work with friends, then theoretically, the painting program was getting in the way. There was no way PaintBerri could ever feature-match popular desktop painting programs like Photoshop and Paint Tool SAI, so we decided to experiment with allowing people to upload their drawings, which turned out to be very controversial.

In the meantime, we sent out a series of surveys based on Lean Customer Development principles and identified 20 extremely loyal artists who I had a 15-20 minute phone conversation with. From them, we learned that interaction with other artists, tangible recognition from peers, and a platform for art role-playing were central themes to our top users.

We also uncovered hard evidence from Google Analytics that told us that artists who participated in groups were more engaged on the site. Almost all the groups on PaintBerri are created to organize art role-plays, stories, and characters.

At this point, we knew that PaintBerri’s value proposition was too diluted, but weren’t sure what direction to focus it in. In a hasty attempt to try and focus the feature set back towards a traditional Oekaki site, we proposed removing the uploading feature, which was met with immediate distress from the community. A subsequent survey about the uploads feature made it resoundingly clear what our community wanted.


PaintBerri leaves its Oekaki roots behind and pivots toward a future as an art role-play and storytelling community

Through months of research and experimentation, the PaintBerri Team finally uncovered what the community wanted: a place to create and share visual stories with friends.

We put up an announcement post detailing the change in direction, and it was enthusiastically received.


Validate assumptions before design and engineering efforts are wasted

The team spent a lot of time listening to loud users instead of the most engaged ones, which led us on a wild feature-building goose chase. We now have a much better understanding of why engaged users are engaged, and can start building in that direction.

It's still too early to tell if the changes will have a positive impact on our key metrics, but now the team is armed with the knowledge, experience, and self-awareness to reduce the chance of future misdirections from happening.

Frances Tung