Reducing busywork for tutor teams

to get the best outcomes for students

My role

Product Designer

User interviews, usability tests, information architecture, UI design, prototyping in Figma, stakeholder alignment, handoff

Team context

I work alongside a fellow Product Designer, under the direction of a Design Lead and in collaboration with a PM.

Status

Currently prepping files for dev handoff

Overview

PLUS Personalized Learning serves nearly 3,000 low-income students with math instruction from 500+ tutors. The PLUS tutor supervisor team loses hours each week to busywork and manual processes. They need an automated, efficient solution to focus on what they do best.

Outcome

My team updated an existing, low-adoption tutor admin dashboard, significantly increasing users’ reported comfort with the PLUS app. We streamlined the information architecture, made screens easier to scan and navigate, and applied the robust PLUS design system.

Jump to screens

User Research

A tutor supervisor is like an air traffic controller.

They monitor multiple tutoring sessions and respond to learning needs as they unfold simultaneously. They also monitor tutor performance over time.

On top of that, tutor supervisors manage schedules, communicate with schools, and implement research trials determined by partners in academia.

“We’re losing half of our data.”

I mapped the sources of all data tutor supervisors take in and put out, to understand how the PLUS data dashboards could better serve them.

There’s no single source of truth for the data related to all of these responsibilities.

One tutor explained that because she’s buried in manual tasks to track down the data she needs, she only receives about half of what she should.

The existing dashboard is just another thing to monitor.

Tutor supervisors reported spending only a couple of hours on the PLUS dashboard per week, mainly for sending mass emails.

Instead, they spent much of their time manually searching for data in Slack messages, spreadsheets, and video recordings.

The existing, low-adoption tutor supervisor dashboard

The Problem

How might we equip supervisors with data to make decisions proactively and predict outcomes?

When tutor supervisors have more time to spend on forward-thinking improvements to tutoring teams, students will receive a higher quality education.

Scoping

How do we balance our dreams against our constraints?

Our dreams:

Tutor supervisors can conduct all of their work in the PLUS app, with no need for other tools such as Slack.

Our constraints:

Time: We need to ship files by the dev team’s deadline for updating our dashboard.
User hesitancy: Our users prefer lots of onboarding information before getting comfortable with major UI changes.

We prioritized design updates that would boost tutor confidence in the PLUS app and improve their day-to-day work. We’re planning for a second phase to build on our improvements with AI.

Phase 1 (shipping now):

Incorporate data visualizations and tutor performance metrics into the dashboard
Align UI with PLUS design system and UX design best practices

A sneak peek at our scoping conversations

Dividing the design work into two phases

Phase 2 (up next):

Explore AI-powered features that will help supervisors with proactive decision-making about the tutors they manage
Update visual design for existing data visualizations (shown in next section)

Information Architecture

Supervisors have specific questions about tutors, and they need answers fast.

We can streamline the supervisor workflow by making info about tutoring sessions easier to find.

Before

Existing dashboard with my annotations from supervisor interviews

After

Updated dashboard with my annotated improvements. Data visualizations were shipped by a prior team and will be updated by my team in the next phase of our work.

Design System

Incorporating the extensive PLUS design system

We leveraged existing patterns when making decisions about usability, efficiency, and aesthetics. Here are just a few examples of how we did it.

Adapting a modal

I adapted the design system’s customizable modal when I was asked to connect our dashboard with specific details about each tutoring session.

Modals as seen in the PLUS design system

My modal for tutoring session details

Replacing a navigation item that didn’t work for our users

The design system provides several types of navigation items, including clickable dropdowns – which our audience didn’t realize were clickable. We decided to reorganize the content so that only one menu item type would be needed.

Before: Supervisors didn’t realize this nav item was clickable.

After: we reorganized content so that we could constrain ourselves to one menu item type from the design system

Choosing a pagination variant

The design system allows for multiple sizes and styles of pagination. In consultation with the design system team, we chose the medium-sized, unlabeled version for cohesiveness with similar dashboards elsewhere in the product.

Pagination variants in the PLUS design system

Pagination as seen in our dashboard

UI Refinement

Mapping all possible states of each table cell

With annotations for developers.

Paring down the data visualizations

Before:

Supervisors feel overwhelmed by the visualizations and immediately scroll past them.

After:

We identified the most useful visualizations for each page in interviews with supervisors. Fewer visualizations, redesigned to take up less vertical space, orient without overwhelming.

Here’s a closer look at my decision-making:

Rewriting confusing labels

How it started…

How it went…

How it ended:

Ensuring responsiveness across 4 breakpoints

Pivot

Supervisor interviews didn’t tell the whole story.

PM feedback revealed gaps in our understanding of the problem, which we addressed in under a week.

During our research interviews, tutor supervisors discussed many issues with how they access data – but didn’t mention that they are responsible for creating much of the data themselves.

We ran a CRUD analysis with the PM to identify where and when supervisors create, read, update, and delete data. As a result of this analysis, I created a new Edit Tutor modal giving supervisors full control of tutor data in one place.