Helping IT teams automate employee offboarding
3 Month Internship
Research, UI/UX, Prototyping, User Testing
1 Designer, 2 PM, 3 Eng
The Problem
The average employee signs up for over 80 apps. Lumos helps IT teams automatically offboard employees from these apps through integrations. My task was to redesign the existing flow and scale it from 15 to 100 newly engineered app integrations.
Outcome
400% increase in customer acquisition with companies such as Codecademy, Khan Academy, Formlabs, View and RapidSOS.
Raised $4.2M.
The challenge
How might we help users setup hundreds of new app integrations?
Research
To define the problem, my research consisted of three parts
1. Audit the existing design to build context
2. Talk with engineers for constraints and technical requirements
3. Interview existing and target users to generate key insights
1. Audit the existing design
Simple at first glance…
In order to build context, I audited the end-to-end process of integrating an app through Lumos. To my surprise, all 15 current app integrations followed a simple process (below) and I was confused about why a redesign was even necessary.
2. Talk with engineers
…but not scalable!
Knowing that this was a highly technical product, I wanted to grasp technical requirements early on by talking with my engineers. My approach paid off! I learned the current designs wouldn’t scale to new technical requirements for integrations with 100 new apps.
Key Insight: Each app has multiple integration methods which automate different processes in the back end. Integrations are unique.
3. Interview target users
Integrations require maintenance
To understand the pains of current and target users, I interviewed 31 chief information officers, IT heads, and leads (thank you to my PMs for helping source participants)! I learned every company offboards employees differently. However, there was a common theme:
Key Insight: Integrations can expire and their credentials must be updated over time!
The goal
Create a scalable and maintainable experience for 100 app integrations
Ideation
I led an ideation workshop with my entire team. By discussing prioritization and feasibility, we explored directions aimed at being scalable and maintainable (our two goals). This process also helped build buy-in and alignment from non-designers early on.
At the same time, I fleshed out information architecture with considerations to a new onboarding flow to be engineered next quarter. One challenge was systems thinking and considering how different onboarding paths could impact the UX in my project’s scope. For example, users who onboard with Google OAuth would not integrate with G Suite again.
Challenge 1
Designing to Scale
From my earlier conversations with engineers, I knew that scaling this experience would largely revolve around new technical requirements:
More Apps: offer integrations for 100 new apps
More Integrations: each app has up to four integration methods
More Automation: each integration automates different processes
Below are decisions from usability tests with 30 target users.
Direction 1
Drawer
My initial hypothesis was to minimize clicks. However, a drawer was not scalable for content in the other tabs. Participants also had difficulty quickly choosing between integration methods (Okta, API, User Impersonation).
Direction 2
Single Page
I improved scalability by using a single page. However, users had trouble drawing a relationship between the affordances of each integration (represented by the checks) and the checkboxes on the right. The numerous CTAs confused users.
Final Direction
Progressive Disclosure
Using progressive disclosure was a tradeoff that added more clicks but reduced complexity. With so many decisions, inputs, and statuses to understand, dividing informative content and actionable content was crucial.
Now, purely informative content (differences between capabilities, statuses and affordances) was separated from user input (adding credentials and choosing actions).
Challenge 2
Designing App Statuses
When an integration breaks and automation in the back end halts, our customers get no value. As previously discovered from interviews, users want to understand exactly what apps, and integration methods and automated actions need maintenance.
Direction 1
Pills
I aimed to simplify statuses with a pill for each integration method. However, this design obscured granularity. For example, capabilities can be broken or connected at the same time under the same integration. Users had no indication of what capabilities needed maintenance.
Validation
Usability Metrics
Throughout these 30 moderated user tests, I tracked key usability metrics to validate my design decisions. This scorecard also served as an additional artifact for earning buyin with high-level stakeholders.
Final Design
Outcome
400% increase in acquisition with customers such as Codecademy, Khan Academy, Formlabs, View and RapidSOS.
I standardized handoff processes by shipping the first 75 screens in the company’s largest UI refresh.
Owned and led the creation of an entirely new design system for our web product.
Raised $4.2M
