GoCo

Senior product designerAcquired by Intuit2022–2025 · SaaS, HR Tech

Three years as one of two product designers on an HR, benefits and payroll platform for 200,000+ people. I designed across more than ten product areas, turned a loose set of components into a design system, and led the rebuild of Reports, the feature customers trusted least.

Role: Senior product designer, 2-person design team
Timeline: Jul 2022 – Apr 2025
Worked with: 8 PMs and multiple engineering teams
Scope: 10+ product areas and the design system

200K+Platform users

+60%Customer satisfaction

↓30%Support tickets

10+Product areas shipped

GoCo employee record with personal details and full employment info

Part one

Three years across the platform

Breadth across the product, the system underneath it, and how the team made decisions.

Overview

HR software where a mistake lands in someone’s paycheck

GoCo serves small and mid-sized U.S. businesses that rarely have an HR department. They have one overloaded person who onboards hires, approves leave, runs payroll and enrolls people in benefits, often in the same afternoon. Every ambiguous screen turns into a support ticket, a payroll correction or a benefits error.

I joined in July 2022 as one of two product designers, partnering with eight product managers across several engineering teams. When a Director of Design joined two years in, I worked with them to change how the team makes decisions. I stayed through the Intuit acquisition in early 2025.

What I shipped

Breadth: more than ten product areas

I wasn’t assigned a feature. I was accountable for the whole surface feeling like one product, which mostly meant making dense admin work predictable.

Scheduling screen — 01Scheduling
A weekly shift calendar with drag-to-assign instead of forms.
*45 → 15 min to schedule a week*

Employee profiles screen — 02Employee profiles
One record for a person’s timeline, notes, reviews and employment changes, replacing five separate screens.

Workflows screen — 03Workflows
One engine for onboarding, background checks and offboarding, with progress and owners visible.

Time off screen — 04Time off
Requests and their context side by side, so approvals don’t need a second screen.

Payroll screen — 05Payroll
Prepare, preview and summary before anything is paid, with bulk actions for large teams.

Workflow builder screen — 06Workflow builder
Admins define steps, roles and permissions for their own processes.

Compensation changes screen — 07Compensation changes
A guided flow for any change that touches pay.

Company and homepage screen — 08Company and homepage
A company hub that routes admins to what needs them today.

Design system

Depth: from a pile of components to a system teams relied on

Components existed, but nothing governed them. Teams designed screen by screen, and twelve modules had drifted into twelve dialects. I built the system (components, tokens, patterns and documentation) and, more importantly, the habit of using it: new work started from the library, and exceptions had to earn their way in.

120+Components built

↓35%Design-to-dev handoff time

12+Product modules unified

0“Which version is right?” questions

GoCo design system from the production component library: buttons, selection, status, menus, fields and date pickers, wizard steps, page headers, table rows, accordions, drawers and tabs

How the team worked

Good screens don’t survive a bad process

With the Director of Design, I replaced ad hoc sign-off with a rhythm the team could repeat.

Critique work in progress

Weekly reviews of unfinished work, not finished screens. Problems surfaced about two weeks earlier.

Evidence before design

No feature entered design without interviews or usage data behind it.

One library

A single Figma library as the source of truth for every team.

One handoff format

Specs, interaction notes and edge cases in one place, so engineering stopped guessing.

Part two

Deep dive: rebuilding Reports

The feature customers trusted least, why it failed, and the three calls that turned it around.

The problem

Reports had stopped being trusted

HR managers use reports to answer questions they’re accountable for: headcount, pay, leave, benefits. GoCo’s builder showed every option on one screen with no preview, so the only way to check a report was to export it and look. When people can’t trust a report, they stop using it and start filing tickets.

Repeat until the numbers look right

Open the builder

Every field, filter and option competes on one screen.

<28%monthly usage, the lowest of any major feature

Configure without feedback

No preview, and filter logic most admins couldn’t explain.

Export to verify

The only way to see a result was to download it.

2–3exports before a report was right

Correct and repeat

Back into the builder, guessing which setting was wrong.

Escalate to support

When the numbers couldn’t be trusted, a ticket was the fallback.

30%of all support tickets

<28%monthly usage, the lowest of any major feature

2–3exports before a report was right

30%of all support tickets

My role

Owned end to end, from research to handoff

I led Reports from the first interview to the handoff spec, with the product manager and the engineers who owned reporting.

What I owned

Research through spec

Interviews, ticket analysis, the competitive audit, the flow, prototypes, usability tests and the spec.

Who I worked with

PM and reporting engineers

Scope and sequencing with product; query feasibility and performance with engineering.

Constraint

Nothing could break

Thousands of saved reports had to open in the new builder without migration.

Constraint

One release to earn trust back

The first release had to kill the export loop, not just look better.

What we learned

The data was fine. The sequence wasn’t.

We spoke to six HR managers, analysed support tickets, and audited how other reporting tools handle complexity.

6 interviews

Trial and error, every time

Every manager described the same loop: guess the filters, export, check, start again.

Support data

Two failure types, 30% of volume

Both were builder problems, not data problems.

Competitive audit

Everyone else reveals options in steps

Looker, Mode and even Google Sheets use progressive disclosure. GoCo showed everything at once.

The insight

The builder asked people to make every decision at once and gave them no way to see the result. The fix wasn’t fewer options. It was the right order, and immediate feedback.

User flow

Before · the old builder, as admins described it

After · the flow we designed

Interview guide

Interview guide · HR admins

01Context

Your role, team size and the reports you run each month

02The last report you built

Walk me through it, start to finish
Where did you stop, or go back?

03Trust

How do you check the numbers are right?
Who do you ask when they look wrong?

04Support

When do you give up and file a ticket?
What would you need to see to trust a report without exporting it?

Pain points

Building3

Every option on one screenFields with near-identical namesFilter logic is unclear

Checking3

No preview before exportExport to check every changeDon’t trust the totals

Giving up3

Easier to email supportSame report rebuilt monthlyCan’t find last month’s report

Rebuilt from my research notes: the admin flow before and after, the interview guide, and pain points from interviews and support tickets, grouped by theme.

Key decisions & trade-offs

Three calls that shaped the new Reports

Decision 01

Sequence the decisions instead of removing them

The call

Fields, then filters, then schedule, each on its own step, with advanced options collapsed until someone asks. Power users keep every capability; new users meet them one at a time.

The trade-off

A wizard was easier to build but locked experienced users into a path they didn’t need. A sidebar kept everything on one screen and left no room for a preview. AI-suggested reports were tempting and too risky for a trust problem. Tabbed steps guide without taking control away.

The report builder, with the step-by-step settings panel highlighted

Decision 02

Pay for the preview in engineering, not in support

The call

Results update live as people configure, so a wrong filter shows up in seconds instead of after an export.

The trade-off

A “dry run” button would have been cheaper, and it would have kept the loop generating the tickets: click, wait, check, go back. I made the case that real-time queries would cost less than the support load, and engineering took on the harder build.

The report builder, with the live data preview highlighted

Decision 03

Defaults over templates

The call

Every report type opens with its most common fields already selected. Time to a first report dropped by more than half.

The trade-off

Templates looked helpful but added a decision before any value, and boxed people in once chosen. Defaults are a starting point people edit, not a choice they have to make.

GoCo reporting overview with standard report packs and custom reports

The new Reports

Reports, end to end

Five real screens from the shipped builder, in the order an admin meets them.

Plays automatically · click any step to jump to it

Before vs. after

What changed for HR managers

Before

After

Every option on one screen→Three focused steps

Export to find mistakes→Live preview while configuring

A blank slate every time→Smart defaults for each report type

Under 28% monthly usage→52% monthly usage

30% of support tickets→Support tickets down 30%

Outcome

From least trusted to fastest growing

↓30%

Support tickets

Reports had been the largest single source.

Before

After

2.2×

Faster report creation

Time to a first report, with defaults and live preview.

28→52%

Monthly active usage

Up from the lowest of any major feature.

+60%

Customer satisfaction

Across the platform during my time there.

The three-step pattern then became GoCo’s standard for complex setup: onboarding, survey creation and benefits enrollment all adopted it. A fix for one feature became how the product handles complexity.

Reflection

Enterprise users want depth. They just want to choose when to reach it.

If I ran Reports again, I’d A/B test the defaults instead of shipping one set. A 20-person company and a 500-person company ask very different questions, and one set of defaults can’t serve both equally.

In their words

Credits

Kyler HolwayProduct Manager

Erin PhillipsProduct Manager

Michael GugelCo-Founder & CPO

Guilherme AngelucciGraphic Designer

more projects

RisevestFintech

Redesigned an investing app used by 620K+ people; onboarding rose from 34% to 58%.

view project →Lenox ParkFintech, DEI analytics

Sole designer on an analytics platform for investors with $4T+ in client assets.

view project →QuickFleetLogistics, Electric mobility

Co-founded an electric delivery network, and designed the product it runs on.

view project →DRO HealthHealth tech, Telehealth

Redesigned telehealth for 50K+ users in Nigeria; sign-up drop-off fell 18%.

view project →