The evolution of Crystal Knows’ onboarding flow
Crystal Knows utilizes a proprietary algorithm to predict individuals' personalities based on their online data. Customers primarily use this information for three purposes:
Understanding how to best sell to someone
Determining whether someone is a good fit during the hiring process
Receiving advice on collaborating effectively with others
The problem
Over 50% of new users weren’t completing the onboarding process.
The product consisted of two key workflows: one in a web app, and one in a Chrome Extension.
This created a fork in onboarding that led to more drop-off.
Understanding our users and their usecases
To create a more effective onboarding experience, we analyzed product data to identify our highest-value users and their core use cases.
Our key question: “Which conversions drive the most impact for our business?”
By examining our organic sign ups, we segmented onboarding into three primary categories:
Sales
Team Management
Hiring
What levers can we pull to increase meaningful engagement from new users?
First I looked for small changes that could be made quickly. I realized that after account creation, users were prompted to self-identify their segment to get the onboarding flow they needed.
Initial approach
Most users were from enterprise accounts and were using Crystal because their manager told them to - they weren’t always sure what segment they were in.
I switched from options based on segment to actual use cases. Users were more likely to know what tool they wanted to use, and offering a way to access the other flows later reduces extrinsic cognitive load.
A large number of new daily users allowed us to track these changes in the wild.
Just by changing the way our onboarding options were presented, our conversion rate increased by about 8%.
At the beginning of a project, I think it’s important to look for the small changes that can produce results. Here we spent 1 week on design and development, rather than a full-scale redesign.
First learnings
Prediction-first onboarding: zero click value
As our technology advanced, the need for an extension used in 3rd party tools became less crucial. This was a really fun opportunity to remove cognitive load from the user, while making the tool feel more “magic”.
Original onboarding flow
Prediction-first onboarding flow
New flow keeps user in one browser tab, no forks where they get lost
Reduces manual inputs from user
Feels more “magic”
Prediction-first onboarding allowed me to treat Crystal’s web app as the “home base” where users could find content and prompts for further advancement.
I created an onboarding task list to Improve the discoverability of our tools. This allowed users to move at their own pace (no one likes feeling stuck in a tutorial pop-up), and aligns with a gamification mindset: humans love to complete tasks and feel like they’ve accomplished something!
Adding a task list
Flow with the addition of the task list
Users are greeted with a confetti pop on first entry to the web app. They’re invited to view their profile, but also given other ways to continue onboarding.
The level up tab acts as a global component that we can trigger notifications from. We did this at time-based increments when users hadn’t completed tasks.
The task list is accessible from any page by clicking the level up tab. Here you see how we continued to encourage them to download the Chrome extension.
I did two cycles of ideation testing for:
Imagery, verbiage, and color to grab the users attention and persuade them to take action
Placements for the task list: in-line as a carousel on the landing page and as a collapsable, global component
Findings
A global component was much more discoverable and didn’t add more load to the already content-dense profile page
The collapsable tab made it easy to hide but allowed us to call it out when additions were made or after a set time
Version testing
How I integrated behaviorism principals through gamification:
Displaying a progress bar and profile strength meter helps motivate
Providing steps the user can take at their own pace lets them feel in control
Using elements that spark curiosity around personality types
Balancing our users goals and gamified elements
Rewarding tasks with free credits acts as a feedback mechanism