Onboarding: the framework
and the redesign

I'd been working on wrist wearable (Gen 1) out-of-box experience with two PDs. We kept getting the same feedback from leads: too long. But we couldn't remove screens. I had a hypothesis — it wasn't the screen count, it was the click count and the narrative momentum — but I couldn't get the PDs to engage with the idea through conversation alone.

So I stopped arguing and started showing. I opened FigJam and built out every possible way to reorganize the same screens: different groupings, different sequences, different splits between what was mandatory and what was optional. Just colored squares. No polish. The point was to make the options visible so we could evaluate them together instead of debating abstractions.

The PDs came around. We collaborated on a version we all had a hand in, and took two variants through UX research.

The underlying concept — mandatory vs. optional out-of-box experience sections — had roots in a conversation I'd had with a previous product design who'd since left the team. We'd called it CHOOBE: choose your own out-of-box experience. The idea was that every team's feature could be included in setup, but not all of them had to be completed by the user. Teams got what they wanted; users got optionality. That's the architecture the CYOA model was built on.

Four flow variants, mapped in a shared FigJam. Each row shows a different way to reorganize the same screens — splitting mandatory setup from optional configuration, reordering features, testing different grouping strategies. No polish. The point was to make the options visible so the team could evaluate them without defending the status quo.

Once the PDs were on board with the concept, we needed to make the case more rigorously — without formal UX research, because we didn't have time. So we ran a structured workshop grounded in jobs to be done and requirements-focused thinking. The goal was to establish shared principles for what belonged in out-of-box experience and what didn't, and to build a framework we could use to defend every decision to engineering and leadership.

We anchored the workshop in the product's "Reasons to Buy/Wear" framework — four jobs the device was hired to do. These became the filter for every out-of-box experience decision: if a screen didn't directly support one of these jobs, it was a candidate for removal or deferral. We also established "Principles for Drawing a Line" — three questions any team could use to evaluate whether their feature belonged in mandatory setup.

We built a journey map for Carla — a "trendy connector" persona — walking her through every step of the out-of-box experience from unboxing to first use. For each step, we documented the task list, her feelings, and the improvement opportunities. The emotion curve at the bottom made visible what the screen count had obscured: the experience had too many low points in the middle, right when users needed momentum.

We mapped the "golden path" — the sequence of moments where users would first feel the device was worth it. These didn't have to happen in order, but they had to happen. The wave diagram made the argument: out-of-box experience should be designed around getting users to their first a-ha moment as fast as possible, not around completing a checklist of feature introductions.

The workshop produced three new user experience principles that became the shared language for every out-of-box experience decision going forward. These weren't just documentation — they were the criteria we used to evaluate the revised flow with engineering and to make the case that we could ship without formal UX research.

The workshop produced a set of revised flow variants — no longer just colored squares, but a structured architecture with clear mandatory and optional sections, color-coded by category. The before/after comparison shows how the flow changed: a flat sequence of screens became a tiered system where the mandatory path is short and the optional path is genuinely optional.

Once the mandatory/optional framework was established, teams had to agree to it — including teams whose features didn't fit the definition. The framework was clear: mandatory setup included only what was necessary for the device to function, or what was a hard requirement from Legal, Privacy, or Consent. Everything else was optional.

One team pushed back. They wanted their feature in mandatory setup. Their argument: their feature was important enough to warrant it. The counter-argument: every team thinks their feature is important enough. That's the whole reason the framework exists.

I held the line. The framework wasn't a suggestion — it was the architecture that made the CYOA model work. If mandatory setup expanded every time a team made a case for their feature, it would collapse back into the same bloated linear flow we'd just fixed. The framework survived. The feature went into optional setup. The model held.

The CHOOBE architecture — mandatory setup separated from optional configuration, with a modular "get more from your device" screen — was adopted across every device in the org. Below are six versions of that screen across different products. The pattern is the same: a clean list of optional setup tasks, each self-contained, each skippable. The content adapts to the device. The architecture doesn't change.

The CYOA work surfaced a broader pattern problem: teams were using double-confirmation dialogs throughout the out-of-box experience. These go against general UX best practices — especially in permissions and consent flows where interrupting momentum erodes trust. I wrote an org-wide guideline to eliminate them, audited every module, and set a deadline for the one team that hadn't complied.

In 2025, I applied the same audit methodology to the wrist wearable out-of-box experience — a 70+ item content audit covering every screen from power-on through the product tour. The audit used a four-category issue taxonomy: missing (content entirely absent), accuracy (incorrect or outdated content), consistency (terminology or tone varies without reason), and quality (content could be clearer or better tailored to the surface).

The audit surfaced patterns that went beyond individual screens: codename accuracy issues throughout, permissions and consent flows that needed privacy team alignment, a product tour that didn't reflect the device's positioning as a companion product, and an etiquette screen that had been carried over from glasses without being reconsidered for the wrist context.

A peer noted in my review that I wasn't sure this was "a battle worth having." They were right. I hesitated. The concept only traveled because I eventually committed to the pitch — but I lost time second-guessing myself first.

The lesson: if you've done the research and the data supports the change, trust your judgment earlier. Stop framing good work as a "bad first draft" to make it feel safer to share. And when words aren't working, build something. The FigJam squares did more in one session than weeks of conversation had.