Marcus Andrey Study
Ler em português Library

UX Research · For leads · 5 min read

Democratising research without losing quality

For leads: there will always be more research to do than researchers to do it. How to let designers, product managers and others run studies well: match the study to the skill, give people guardrails instead of gates, coach rather than police, and check the results.

The question isn't whether people outside research will talk to users. They already do. The question is whether they'll do it well. A product manager who runs a leading interview and a designer who tests with colleagues aren't the problem; nobody showing them how is. Democratising research means making the good way the easy way, and keeping the hard studies for people trained to run them.

What you'll be able to do after reading:

  • Explain why democratising research is worth doing, and what can go wrong
  • Match each kind of study to the skill it needs
  • Put guardrails in place that raise quality without slowing people down
  • Check whether it's working

1. Why, and the risk

Steve Portigal states the problem plainly: "The demand for User Research exceeds the supply of researchers." His answer is that research is a helping profession, and "we should be finding ways to empower everyone to do well" (Portigal, 2020).

NN/g agrees, with a condition. "Even imperfect user research is usually much better than no research at all," writes Kara Pernice, but without assessment, training, coaching and clear accountability, democratising leads to poor research, poor designs and wasted effort. Her warning: "Don't get caught in the trap of believing that people without training, experience, or knowledge can do research well" (Pernice, NN/g, 2022). Both are true at once, and the lead's job is to hold both.


2. Match the study to the skill

Basic skills

Anyone on the team, with the kit

Usability tests of a prototype, with a template task list. Short feedback interviews on a shipped feature. Observing and taking notes.

Support: templates, a plan review, one coached session.

Intermediate skills

Trained, with coaching

Discovery interviews on a known topic. Surveys that inform, not decide. Synthesis of a small study.

Support: training, office hours, an experienced person in the synthesis.

Advanced skills

Experienced researchers

Foundational research that sets strategy. Sensitive topics or vulnerable participants. Surveys and benchmarks that decide. Mixed-method studies.

Support: they coach the other two columns.

Lower risk if it goes wrong · reversible decisionsHigher risk · strategy, people's safety, big bets

Who can run which study, by the skill it needs and the cost of getting it wrong. The line moves as people grow: the point is that everyone knows where it is today.Diagram: Marcus Andrey, CC BY 4.0. Skill levels after Kara Pernice, Democratize User Research in 5 Steps (NN/g, 2022); the examples in each column are this course's.

Pernice's model is practical: classify who wants to do research and their skills as basic, intermediate or advanced, give simpler work to people with less experience, and keep experts for complex studies, with coaching in between (NN/g, 2022). The figure is one way to draw that line. Where exactly you draw it depends on your people, not on their job titles.


3. Guardrails, not gates

A gate stops people until someone approves. A guardrail lets them go and keeps them on the road. Good guardrails:

  • A starter kit: plan, screener, guide and report templates, a consent form, a retention rule (see ResearchOps).
  • A 15-minute plan review with someone experienced, before recruiting starts. It catches the wrong question, the leading guide and the missing consent while they're cheap to fix (see Write a research plan).
  • Office hours where researchers coach, a format NN/g recommends (NN/g, 2022).
  • Shared synthesis: the study's owner runs it, someone with experience joins (see Synthesise a study).
  • Everything in the repository, labelled with who ran it and how, so others can judge the evidence (see Research repositories).

4. Coach the skills that go wrong most

In practice, the same few things go wrong in studies run by people new to research, and they're all coachable:

Sit in on one session per new practitioner, and give feedback in private, specifically, the same day.


5. Check that it's working

Pernice's last step is to revise: look at how much demand was met, whether designs improved, whether teams embrace research, and whether there are enough coaches (NN/g, 2022). Her model for judging the research itself is V.I.S.E.: research that is viable, influential, sound and efficient. A simple review every quarter: pick three studies run outside the research team, read the plan, the guide and the report, and ask whether each was sound enough to act on. If one wasn't, that's a coaching topic, not a reason to close the door.


6. Common mistakes

  • Banning research by non-researchers, so it happens anyway, unseen.
  • Letting anyone run any study, including sensitive topics and research that sets strategy.
  • Training once and never following up.
  • Reviewing only the final report, when the problems were in the plan.
  • Treating coaching as extra work, unplanned and unrecognised.

7. Try it (no tools needed)

  1. List the people outside research who talked to users in the last quarter.
  2. Next to each, write what kind of study it was and how it went.
  3. For each, write the one guardrail that would have helped most.

Reflection questions: Which kind of study would you never hand over, and why? Who on your team could run office hours, and what would they stop doing to make the time?


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.

Building a research practice · Showing the impact of research · ResearchOps · Short user interviews

Saved only on this device.

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