Kate Syuma on Why Miro's Best Onboarding Experiment Failed Users Loved

The Onboarding Flow That Twitter Loved But Users Hated

When Kate Syuma joined Miro in 2017 (then called Realtime Board), there was no user onboarding. Not a broken flow. Not a clunky experience. Nothing.

As the first product designer on the growth team, she built one of the early signup flows: a goal-oriented questionnaire with animated product previews on the side. Users could select their use case—Kanban, mind mapping, wireframes—and watch GIFs of templates come to life. Design Twitter went wild. "Love this onboarding flow," one user tweeted.

The metrics told a different story. Conversion dropped.

After launching that, we saw some drops and then we analyzed why it's happening with user interviews and uncovered that users are distracted by these visuals. They start interacting with them, especially if they look like the product itself.

— Kate Syuma

The lesson: anything that looks interactive in a signup flow will get clicked. And every click that's not moving users toward activation is friction. The team stripped it back—clean, centered, minimal. No animations. Just progress.

What Changed When Non-Tech-Savvy Users Flooded In

COVID changed everything. Suddenly Miro wasn't just for product teams and designers. Teachers, HR managers, consultants—people who'd never used a visual collaboration tool—started signing up by the millions.

The old onboarding assumed baseline product literacy. The new users didn't have it. So the team experimented with something radical for a self-serve product: human touch at scale.

They introduced a "collaborative robot"—a semi-automated guide that mimicked the experience of a customer success manager walking someone through their first board. The qualitative feedback was glowing. But there was a problem: video didn't scale. Not in 2020.

The team turned to Synthesia, an AI video generator, to prototype the concept before building it in-house. It worked for user testing, but the output quality wasn't good enough for production. Even today, Kate says the tech isn't quite there for a quality-obsessed product like Miro.

I still believe today, especially with the power of AI, we should continue experimenting with that direction.

— Kate Syuma

Don't Be Afraid to Add More Questions

Most teams assume more questions = lower conversion. Kate's research says otherwise.

Profiling questions—role, use case, team size—might drop signup conversion by around 10%. But they don't hurt activation. In fact, they often improve it, because users become more intentional when they're asked to articulate their goal upfront.

The key is progressive disclosure and clean UX. Don't throw six questions on one screen. But don't be scared of gathering the data you need to personalize what happens next.

And don't treat profiling as static. Your ICP shifts. Your product evolves. The questions you asked in 2020 might not make sense in 2024. Kate recommends revisiting them at least quarterly.

At 50 Million Users, the Big Bets Get Smaller

When you're at scale, redesigning onboarding from scratch is expensive and slow. The learning curve on a three-month rebuild is brutal. So Miro's growth team stopped chasing big bets and started hunting for high-leverage micro-improvements.

One example: the "say hi" reaction feature. When a new user joined a shared board, they'd see a prompt to wave hello. It introduced them to reactions (a core collaboration feature) and created a delightful first interaction. Small change. Measurable impact.

You start being very intentional about what actually can move the needle and you start searching for that very, very tactical sometimes thing, but the thing that can actually move this needle.

— Kate Syuma

The mistake most teams make at this stage? Overwhelming users with feature tours. Kate's onboarding research—covering more than 80 companies—found that bombarding users with notifications about dashboards, settings, and integrations kills activation. Show them one or two delightful, high-value features in session one. Save the rest.

Three Things You Should Do This Quarter

Kate's advice for any team working on activation:

Review your end-to-end onboarding flow every quarter. Sit down with your PM, designer, and engineer. Find bugs. Spot UX improvements. Build a backlog. Activated users convert to paid at 3 to 10 times the rate of non-activated users. This isn't a nice-to-have.

Audit your profiling questions. Are you actually using the data you collect? If not, replace those questions with ones that help you identify ICP or personalize the experience. If you're not asking profiling questions at all, start now.

Define your activation metric—even if it's rough. It might take six months to get it right. You'll refine it over time. But if you're iterating on onboarding without a clear definition of what "activated" means, you're flying blind.

Kate's building a course on onboarding and publishing more research at onboard.report. She also hosts the GrowthMates podcast, where she talks to growth leaders from Dropbox, Amplitude, Coda, and Canva about what actually works.