← BackNext project

Design & Research Lead · Capital One

Leaders and associates at Capital One distrusted 360 feedback they received. I used research to turn a noisy process into clearer, more actionable growth conversations.

My role

Design and research lead

Timeline

Capital One · Dec 2023 – Sept 2024

Platform

Web platform

Directional signal

0%gained clarity on development opportunities
0%more feedback used in performance conversations
Where the 360 feedback experience broke down across the performance cycle
Performance management end-to-end process. Highlighted areas represent selected scope.

The Problem

Low trust in a system that was supposed to help people grow

360 feedback was poorly connected to the broader performance flow. Feedback templates varied wildly across teams. Responses skewed positive, not because everyone was performing exceptionally, but because the system gave people no reason to be specific or honest. People leaders lacked confidence in the feedback they received. Associates didn't know how it would be used. The result was a process that consumed time and produced noise.

We interviewed leaders and associates and looked closely at how feedback was actually being written and read. The same patterns kept surfacing: people discounted feedback they couldn't put in context, and they softened their input when they weren't sure how it would be used or who would see it.

Detail of the 360 feedback experience breakdown

01

Foundational principles

From research, we knew that we needed to improve the consistency, quality & actionability of the feedback received for it to be useful during the performance process. Our hypotheses were:

Quant & qual data together. Standardized, competency-based ratings paired with required qualitative comments improved consistency while gathering feedback.
Psychological safety through anonymity. Complete anonymity encouraged more candid and constructive responses, improving the quality of feedback received by people leaders.
Comparative context to reduce bias. Use of a comparative scale (e.g., “compared to peers”) reduced bias and subjective ratings, increasing actionability during live calibrations.
Quant & qual data together
Psychological safety through anonymity
Comparative context to reduce bias

02

Connecting feedback to calibration

Before we built the system, we wanted to pilot our hypotheses. We partnered with PwC to build the feedback system on these foundations, grounding every question in Capital One's competency framework and making the entire process anonymous by design.

Feedback form: competency-based ratings, required qualitative comments, fully anonymous
Feedback form: competency-based ratings, required qualitative comments, fully anonymous

The bet wasn't obviously safe. Full anonymity could have made leaders trust the feedback less (it's easy to dismiss a critique you can't attribute), and “compared to peers” framing risked turning a growth tool into a ranking. We were trading those risks for candor, and wouldn't know which way it broke until the pilot.

The key decision: making 360 feedback a first-class input in calibration, not an afterthought. At the time, people leaders used Google Slides to represent their associates during calibrations. For our pilot group, we redesigned the calibration slide to surface feedback directly alongside the performance data leaders collected. Peer comparison graphs showed ratings relative to the cohort. Written feedback was structured to surface strengths and development opportunities side by side, with context on who provided the feedback: something managers could actually reference mid-conversation.

Calibration one-pager: 360 feedback as first-class input with peer comparison graph and written feedback
Calibration one-pager: feedback as a first-class input, not an afterthought

03

Measuring what mattered

After the performance cycle, we measured impact by surveying, observing, and interviewing different participating user groups, then triangulating those data sources. This helped us understand what was resonating with users at each step of the performance cycle, how much the feedback was actually used, and how it shaped performance conversations.

Measurement data, detail 2
Measurement data, detail 3
Measurement data, detail 4
Measurement data, detail 5

04

The pilot made the case

The results were strong enough to convince our HR stakeholders to discontinue using Workday as the primary tool for performance and talent, and invest in building an in-house performance system that understood Capital One's internal performance process and was grounded in 360 feedback as the foundation.

0%

improvement in clarity & consistency of feedback received

0%

improvement in feedback quality: anonymity made a measurable difference

0%

improvement in actionability: feedback used more actively in live calibrations

Directional figures from the pilot study: the signal that convinced HR to invest in an in-house platform.

That in-house platform became PATH (the next case study), where this pilot's bet got built for the whole enterprise.

05

Growth as a designer

This was my first major lead effort, and it changed the way I think about product and strategy design.

What stayed with me most was how much stronger the work became when alignment happened early. Bringing cross-functional partners in from the beginning didn't just improve the solution: it created a shared sense of ownership that carried the project forward. It also taught me that measurement isn't something you do after launch; it's how you understand whether the work is resonating, and how you earn the next phase.

It also gave me a clearer sense of where I'd grow next. We took on a lot of change at once, and in hindsight I'd be more intentional about managing scope, thin-slicing the problem, and sequencing bigger bets so the impact of each decision can be seen more clearly.

That pilot proved better feedback could change the conversation. PATH asked the bigger question: how do you scale that trust across the enterprise?