I built the first research practice for a cloud platform from more than twenty interviews and three hand-drawn charts, and the program ran for nearly five years.

I joined a fast-growing cloud infrastructure team as its first dedicated researcher. I listened to the people who built the product, drew the plan with them on large charts, and turned those charts into a research process, a way to request and triage studies, a team skills map, and a first-year roadmap for the product teams.

  • Client

    A global cloud provider

    I changed names and details to respect my nondisclosure agreements.

  • My role

    Head of research, later research program manager

  • Focus

    Building a research practice from zero

  • Result

    A research program that ran from 2017 to 2022

No map yet

A fast-growing cloud team wanted research, but nobody agreed on the purpose of research.

In August 2017 I joined the design team of a cloud infrastructure platform that was growing faster than anyone could track. The team had talented designers and product managers, and it had no dedicated researcher. Studies had happened now and then, and several designers told me they had never watched a real customer use the console they designed.

Everyone wanted research, and everyone pictured something different. Some people wanted quick checks on a design. Others wanted a deep understanding of customers who were moving to the cloud for the first time. Several worried that research would slow the work down or arrive too late to matter. Underneath all of that sat practical blockers: no approved way to recruit and pay participants, no shared tools, and no place to keep what we learned.

Listen, then draw

I listened to the people who built the product, then drew the practice with them on the wall.

I wanted to understand the team before I proposed anything, and every step below fed the plan. Click or tap a card to turn it over.

What the room told me

The team wanted research that helped them move faster and never stood in their way.

The interviews were candid, and the same few hopes came up again and again. I kept the notes private, and I turned each hope into a rule for how the practice would work.

Those rules shaped the process, the hub, and the way every study ended.

  • Research had to speed the team up, so every request went through a quick triage that said what to study and what to skip.
  • Designers wanted to see customers for themselves, so every study saved seats for a second moderator and a note taker.
  • The people who bought the product were not the people who used it, so the practice studied buyers and users separately.
  • Findings had to be short and easy to find, so every study ended with a summary stored in one shared hub.
From request to result

I designed a process that any team could use to ask for research.

The process model gave every request the same four steps, so nobody had to guess what happened after they asked. Click or tap a card to turn it over.

My opening line

Humans minus technology equals humans. Technology minus humans equals absolutely nothing.

Ashley KarrFrom my first research plan for the team

Try it yourself

Click through the research hub I designed the practice around.

I rebuilt the plan, the request and triage process, the skills snapshot, and the growth matrix as one working hub under the name Haven Cloud, with AI writing the code from my designs. The team members and requests are sample content.

Things to try

  1. Check off items in the plan and watch the progress bars grow.
  2. Open Request research, change where the product is today or when you need an answer, and watch the triage route change.
  3. Send a request, then move it forward through research, closing, and shared.
  4. Open Team skills, choose a person, and change a rating to see a skill gap open or close.
  5. Open Growth matrix, pick up a skill, and move it to another square.
The kit behind it

Olive, mustard, and warm paper carried the colors of the hand-drawn charts into the hub.

I pulled this style guide from the prototype. The palette comes from the watercolor charts I drew with the team, so the hub feels like the same conversation. Every text color passes WCAG AA contrast.

Color

  • Forest#2B3514The top bar and the navigation.
  • Olive#4A5A23Primary buttons, checked items, progress bars, and filled skill dots.
  • Mustard#E0A82EThe current tab, the logo mark, and stage numbers.
  • Mustard ink#6E4C0CLabels on mustard tints, such as the In triage status.
  • Sky#2F5F8AFocus outlines and the In research status.
  • Brick#A63D2CSkill gaps that need a hire or a mentor.
  • Paper#FAF6ECThe warm page background behind every card.

Type

  • Headings

    Tell us what you need to learn

    Bricolage Grotesque Bold

  • Labels and buttons

    Send request

    Figtree Bold

  • Body

    Write a question, not a solution.

    Figtree

Components

  • Plan columns

    Each period of the plan gets a column with its own progress bar, so the team can see how far the first year has come at a glance.

  • Stage picker

    Four numbered stages show the question, the methods, and the measures that fit each moment in a product's life.

  • Live triage panel

    The panel beside the request form updates with every answer and explains in plain words why it recommends one route over the others.

  • Skills table

    Filled dots show each person's own rating, and the last column flags any skill where fewer than two people feel strong.

  • Growth matrix

    Four squares sort each person's skills into good or bad and love or hate, and each square says what the team should do next.

Principles

  • Ask for the question before anyone proposes a solution.
  • Give every request a route, even when the route is a short conversation.
  • Let people rate their own skills, and use the ratings to staff projects, never to grade people.
  • End every study in one shared place so findings outlive the meeting.
Where it went

20+

interviews with product, engineering, and design leaders before I wrote the plan

4

research stages in one shared process, from discovery to optimization and growth

What changed for people

The practice grew from a plan on the wall into a program the product teams used for years.

The program kept running from my first month in 2017 until I left for my next role in 2022. Click or tap a card to turn it over.

If you are starting from zero

A new research practice starts with listening and a picture everyone can change.

The charts mattered because the team drew them with me. People saw their own words on the wall, argued about them, and left with a plan they owned. If your team is starting research from scratch, I can run the same interviews, charts, and first-year plan with you, and build the request and triage process your teams will use every week.

Still curious?

There is plenty more where this came from.

Names, brands, and data in my case studies are changed to protect my clients.

Connect

Tell me about your project.

Tell me a little about it on my contact form, and a real person will get back to you soon.