Marcus Andrey Study
Ler em português Library

UX Research · For leads · 4 min read

Building a research practice

For leads: how to turn occasional studies into a practice the team relies on. Read your organisation's maturity, start from real demand, make research safe and visible before you make it big, and set a rhythm the team can keep.

A research practice isn't a team of researchers. It's a team that keeps learning from people, on purpose, whether or not anyone has "researcher" in their title. Most practices start as one person doing studies when there's time. Leads turn that into a rhythm: research that happens before decisions, that everyone sees, and that leaves something behind for the next study.

What you'll be able to do after reading:

  • Read your organisation's UX maturity and pick a realistic next step
  • Start from the research the team already asks for
  • Build the safe, visible basics before the big ones
  • Set a research rhythm the team can keep

1. Know where you're starting

NN/g describes six levels of UX maturity, from absent and limited (UX happens occasionally, with no dedicated roles or budget) through emergent and structured to integrated and user-driven, where research guides strategy at every level. It measures them across four factors: strategy, culture, process and outcomes (Kara Pernice, Sarah Gibbons, Kate Moran and Kathryn Whitenton, NN/g, 2021).

The level tells you what to build next. In a limited organisation, the win is one study that changes one visible decision. In a structured one, it's consistency: the same consent, templates and repository for every team. Planning a user-driven practice from a limited start is how leads burn out.


2. A first year, in four moves

  1. 1Listen

    List the decisions coming up and the questions people keep asking. Rank them.

    Leaves behind: a research roadmap
  2. 2Make it safe

    One consent form, one retention rule, templates for plan, screener, guide and report.

    Leaves behind: a starter kit
  3. 3Make it visible

    Small rounds every couple of weeks, observers in every round, a show-and-tell slot.

    Leaves behind: people who have seen users
  4. 4Make it count

    Track which recommendations were adopted, and what changed for users and the business.

    Leaves behind: the case for the next year
A first year for a research practice, in four moves. Each one leaves something the next can build on, so the practice grows like a village, not a blueprint.Diagram: Marcus Andrey, CC BY 4.0. After NN/g (Pernice, 2022), the GOV.UK Service Manual and Kate Towsey (NN/g, 2025).

Kate Towsey's advice for operations applies to the whole practice: start small, but map the whole journey first, and tie every system to what the business needs, not to the request in front of you. "Every city started as a village" (Kate Moran and Maria Rosala, NN/g, 2025).


3. Start from demand

Before you plan studies, list what the team already needs to know. NN/g's first step for scaling research is the same: inventory completed studies, unmet requests and both strategic and reactive needs, and rank them (must do, should do, if time, defer, don't do) (Kara Pernice, NN/g, 2022). Two things come out of the list: which decisions are being made without evidence, and which questions keep coming back. The first is your roadmap. The second is your case for a repository (see Research repositories).


4. Make it visible

People trust research they've seen. The GOV.UK Service Manual sets a clear bar: everyone on the team should observe at least two hours of research every six weeks, and teams should plan at least one round of research every two weeks, usually with four to eight participants (GOV.UK Service Manual, Plan user research for your service). You don't need to match a government digital team on day one, but the shape is right: small, frequent rounds, and a team that watches.

Practical ways to get there:


5. Make it safe and repeatable

The basics that every study needs, so nobody rebuilds them: a consent form, a data retention rule, a participant list, templates for the plan, screener, guide and report, and a repository with an owner. These are the eight pillars in miniature (see ResearchOps). They're also what makes it possible to let more people do research without losing quality, the next lesson.


6. Common mistakes

  • Starting with a tool or a big foundational study, instead of a decision.
  • Doing every study yourself, so the practice stops when you're on holiday.
  • Research nobody outside the design team ever watches.
  • No consent or retention rules until something goes wrong.
  • Measuring the number of studies instead of what changed (see Showing the impact of research).

7. Try it (no tools needed)

  1. List the last ten product decisions your team made.
  2. Mark each: research informed it, research came too late, or no research.
  3. For the "too late" ones, write when the research would have had to land.

Reflection questions: At which maturity level is your organisation, honestly? Who outside the design team watched a research session in the last six weeks?


Checklist

Learn more

Books, courses and videos to go deeper

Book links go to Amazon. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.

ResearchOps · Democratising research without losing quality · What research is for · Who does what on a product team

Saved only on this device.

↑ ↓ to move · Enter to open · Esc to close