SIMnet Assignment Manager
For 20,000+ instructors each semester, relying on customer specialists was the only way they knew to handle 2M+ student assignments. I redesigned assignment management, from the architecture down to the interactions, around how instructors actually think so that they can manage assignments on their own.
Highlights
Task completion
From 20% to 90%
Specialist support requests
Reduced by 40%
Org-wide Adoption
9 designers & 4 products
Team
Product Owner, 4 Developers, A11y Coach
Role
Product Designer, end to end
Timeline
Fall 2025
Context
SIMnet is McGraw Hill’s learning platform for Microsoft Office, and assignment management is where instructors live and breathe. Despite being the most-used feature in the product, it consistently ranked last in NPS.
Instructors relied on specialists for everyday tasks
Features existed, but nobody could find them
When I talked with customer success specialists, I kept hearing the same story. Instructors regularly reach out needing help with tasks like updating multiple due dates and reordering assignments. The response was always the same:
That feature exists, but it's located somewhere else.
Only 20% of instructors could organize and bulk-edit without help
Across 10 usability sessions and 7 instructor interviews, the majority struggled to activate the right actions or understand what was available, even when nudged to the right workspace.
Instructors like Pam resorted to tedious workarounds
Even when instructors successfully completed tasks, they weren't always efficient. In a session with Pam, for example, she resorted to editing assignments one at a time just to get things done.
Instructors hit two walls: finding features and using them
Discovers one workspace, usually whichever they're shown first
Assumes that's all there is
Hits a task it can't do, unsure where else to look
Calls support for a feature they can't find
Eventually navigates to the right workspace
Still can't complete task, actions are buried
Calls support for a feature they can't figure out
Tools across workspaces were effectively invisible
Assignment management spans 3 distinct workspaces, but instructors operate with a single-context mental model. An instructor in the organize workspace, for example, had no idea they could bulk-edit in the edit workspace.
Interacting with core features didn't match expectations either
Actions like bulk-editing required a drag-click selection that was easy to miss, and the button to trigger it stayed hidden until you did. Organizing, too, didn't have drag handles or clear indications how nested folders behave. In both cases, instructors had no way of knowing what to do or expect.
How might we help instructors find and use assignment management features on their own?
One workspace, built around how instructors think
All features, with no context switching
I unified everything into a tree grid. The same pattern is used in Jira, Canvas, and Google Drive, with folders always visible and every action reachable without switching contexts.
Dragging is expected to reorder, not select and take action
Instructors expect drag to reorder, so I used checkboxes for selections and bulk edits instead. I moved bulk actions to the top left, where half of our core actions already lived.
Accessibility team vs. product requirements
Our a11y team and product owner couldn't agree on a single solution because accessible references and pre-built grids each had gaps that made them fall short.
This was an org-wide issue where teams were cutting features from roadmaps and removing drag interactions entirely to only re-order with one-click solutions.
To break the deadlock, I learned both sides
I dove into WCAG 2.2 AA criteria alongside our product requirements, understanding both sides well enough to find where they could meet. Accessibility annotations usually happen near delivery, but I moved them earlier so decisions could be made with both sides at the table.
Mapping edge cases with developers and accessibility coach
I considered removing nested folders to align with Canvas and reduce edge cases, but they had heavy use with instructors and were a market differentiator. We kept them and mapped every edge case.
Testing the new concept revealed one more spot to iron out
100% of instructors instantly found assignment tools
Bulk-editing, organizing, assigning and viewing were all clear.
One workspace was easier to understand than three
The unified layout eliminated confusion from context switching.
Having bulk actions and organizing together was too much
Instructors struggled having both checkboxes and drag handles at one time. Most of them assumed checkboxes would control re-ordering.
Discoverability wasn't about separate spaces, it was about context switching
I tried to understand why and found out instructors usually organize at the beginning of a semester when they're setting their schedule up. I pitched the idea of having a separate yet identical mode for organizing.
The final solution: two modes, one workspace
Stakeholders agreed, so I moved organize into its own space. An identical layout with drag handles instead of checkboxes made the transition feel familiar rather than learning something new. This removed confusion from them being together.
Every metric improved
Task completion jumped 20% → 90%
The majority of Instructors could complete core tasks like bulk-scheduling and organizing.
Assignment NPS went up by 6 points
NPS increased from 23 to 29, and all 10 instructors preferred the new experience.
What instructors said
"First, I really liked the look of it and the ease of making assignments. Changing dates in a group was so much easier."
9 designers & 4 products adopted my patterns
The accessible tree grid interaction patterns solved my team's deadlock, and 9 other designers on 4 different products used it as reference.
Instincts are a starting point, not a conclusion
I was convinced instructors wanted everything in one workspace; it's how I'd want it, and how anyone comfortable with modern software tools would expect it to work.
Testing with instructors, however, told a different story, and I had to stay curious enough to let the data lead. The best solution wasn't the one I started with or the one I validated in the middle. It was the one testing found for us.
Know your product partners as well as you know your users
When they couldn't agree, teams across the org were cutting features rather than solving the problem. I read WCAG 2.2 AA directly and worked with my product owner to shape what the requirements actually needed to find a middle ground.
The best tools disappear
I don't have a teaching background, so every interview was a reminder that the people using this product have a full job that has nothing to do with it. The bar isn't "can they learn it," it's, "can they use it without thinking about it."