Genius AI
Growth for GlossGenius
CONTEXT
Optimizing onboarding in the middle of a rebrand
The Growth team at Genius AI was in an interesting spot. Our charter was to improve growth metrics, like onboarding, activation, and retention. At the same time, our company was in the middle of a high-profile rebrand, and we owned high-visibility surfaces like onboarding and pricing pages. Thus, my job was not only to strengthen performance, but to bring visual delight and brand alignment to these surfaces.
CHALLENGE #1
"Make it look AI"
Exec was invested in branding the new "AI era" of our GlossGenius product, a booking and payments platform for beauty business owners. This meant that leadership wanted to turn our onboarding survey into a chat format.
However, I hypothesized that users wanted to explore the product first - and that the extra time waiting for the chat to "respond" could actually hinder their progress. The data showed that each survey step had a 97-99% completion rate, which I took as signal that users would rather speed through the survey to explore the product.
Before (left) and after (right)
I pushed back on the chat idea, and through exploration, landed on compromise - we kept the survey, but make it seem "dynamic" by displaying a relevant, snappy "response" in the header section whenever a choice was selected. Users were not blocked from proceeding if the "response" hadn't finished loading.
Ultimately, the impact was no harm to survey completion or D14 activations.
CHALLENGE #2
Service setup was buried and inefficient
Impact:
5.99% lift in D1 service setup, from 49.35% -> 52.30%
Matured out to +4.91% by D7.
One key onboarding step was for small business owners to add a menu of services that they offered.
Before users could add services, they had to go through a flow that asked for their business address, hours, and booking settings. The settings were complex and difficult to understand at a glance, and many users were dropping off before getting to the most important task (adding a service).
Before:
I bundled a few high-confidence bets together:
Moved service setup to the beginning of the flow, and brought suggested services front
and center.Implemented smarter settings defaults based on data.
Simplified 3-way decisions into cascading binaries.
After:
CHALLENGE #3
Testing out the platform
Impact: 31.61% increase in booked appointments, from 20.25% -> 26.65%
Now that a user has set up their services and booing rules Book a test appointment and make a test checkout to learn about the scheduling and payment.
These were launched as 2 separate experiments. The "test checkout" experiment still in progress, check back for results!
After service setup: new test appointment experience
New test checkout experience:
CHALLENGE #4
Making sure people finish onboarding
Impact: Experiment still running, directionally positive on D14 activations, stat sig positive on D1 payments verification (8.77% lift) and D1 payout account setup (11.17% lift).
Users who choose to skip onboarding steps see them surfaced again on the dashboard. Previously, these tasks sat underneath the main section of the. I moved the main tasks to to the top, into the main spotlight.
Before (left) and after (right)
Secondary metrics
IMPACT
Learnings and Takeaways
Over the quarter, I was able to completely transform the onboarding experience. While the onboarding tasks were previously disjoint, there is now a series of walkthrough demos that are tied into each other. Additionally, I eliminated obvious friction with bundled bets.
Sometimes there's a reticence to "blow up the flow," but with each extra "demonstrative" step we added, we actually saw an increase in engagement and activation metrics. In the future, I want to fold in even more product education into these demo experiences.
By removing "bad friction" (choice overload, buying the lead, no or bad defaults) and adding "good friction" (education, demonstration, explanation), I was able to make an impact on growth!