Reimagining Onboarding at Canopy

How we reduced time-to-value from 12 minutes to 3.

Onboarding flow wireframes — version 7

The Problem

Canopy is a B2B SaaS platform that helps small marketing teams automate their campaign workflows. When I joined as the lead product designer, the onboarding experience was a 12-minute gauntlet of configuration screens. Users had to set up 8 different parameters — team size, industry vertical, integration preferences, notification rules, billing tier, timezone, workflow templates, and approval chains — before they could see a single dashboard.

The numbers told a painful story. Analytics showed a 40% drop-off rate at step 3 (integration preferences), and only 58% of users who started onboarding ever completed it. Support tickets about "how do I get started" outnumbered every other category combined. Users were churning before they ever experienced the product's core value.

The CEO flagged onboarding as the company's top priority after a board meeting where investors questioned our activation metrics. The pressure was real — we had about six weeks to show measurable improvement before the next funding review. I knew we couldn't just polish the existing flow. We needed to rethink what onboarding meant entirely.

Research

I scheduled 15 user interviews over two weeks, deliberately mixing recent signups who completed onboarding with those who abandoned it. My go-to technique is deceptively simple: "Walk me through the last time you tried to set up a new tool for your team." I never mention Canopy specifically at first — I want to understand their mental model of onboarding in general before narrowing in on our product.

The pattern was unmistakable. Eleven of fifteen participants said some version of the same thing: "I just wanted to see if it could do the thing I needed. I didn’t want to configure anything yet." They weren’t resistant to setup — they were resistant to setup before proof. They wanted to see one working example of Canopy automating a campaign before investing time in configuration. The integrations screen (our biggest drop-off point) felt especially premature — users didn’t know which integrations they’d need because they hadn’t seen the product work yet.

The Solution

We designed a progressive disclosure onboarding model. Instead of asking users to configure everything upfront, we'd show them a pre-configured workspace immediately — populated with sample data from their industry vertical (which we could infer from their email domain and one simple question). The required fields dropped from 8 to just 2: name and email. That's it. Everything else became optional, surfaced contextually as users explored.

The first thing a new user saw after entering their name was a working campaign dashboard with realistic sample data. A gentle tooltip said: "This is what your campaigns could look like. Let's make it real." From there, each configuration option appeared only when the user encountered a feature that needed it. Want to connect your email provider? That prompt shows up when you first try to send a campaign — not during onboarding.

The hardest part wasn't the design — it was convincing the engineering team that generating realistic sample data for each industry vertical was worth the investment. We built 6 industry templates (e-commerce, SaaS, media, non-profit, education, and agency) and wrote a lightweight inference engine that matched new users to templates based on their email domain and a single dropdown: "What best describes your team?"

Before vs. after — 8 required fields reduced to 2

Results