Pick a research method by business task, product phase, or AI compatibility. Click any method to see details and a ready-to-use AI prompt.
Showing all 40 methods
Understand
1
Di
In-depth Interview
2
Ci
Contextual Inquiry
3
Et
Ethnography
4
Ds
Diary Study
5
Js
JTBD Switch
6
Si
Stakeholder Interview
7
Fg
Focus Group
8
Ps
Personas
9
Cj
Christensen JTBD
Generate
10
Pd
Participatory Design
11
Jm
Journey Map
12
Sb
Service Blueprint
13
Mm
Mental Models
14
Cs
Card Sorting
15
Ct
Concept Testing
16
Jc
JTBD Canvas
Validate
17
Ut
Usability (Moderated)
18
Ur
Usability (Unmoderated)
19
Ab
A/B Testing
20
Tt
Tree Testing
21
Fc
First Click Test
22
Dy
Desirability
23
At
A11y Testing
24
Bm
Benchmarking
Measure
25
Sv
Survey
26
Np
NPS / CSAT / SUS
27
An
Analytics
28
Fa
Funnel Analysis
29
Co
Cohort Analysis
30
Hm
Heatmaps
31
Mx
MaxDiff
32
Ka
Kano Model
33
Od
ODI Research
34
Ti
True Intent
Analyze
35
Dr
Desk Research
36
Ca
Competitive Analysis
37
He
Heuristic Eval
38
Er
Expert Review
39
Cw
Cognitive Walkthrough
40
Lr
Literature Review
Understand Generate Validate Measure Analyze
Research designs
Launch a new product
You're about to build something new and need to validate the market before spending engineering time. Works best when the category exists but your position in it does not.
Method chain
Expected deliverables
A market synthesis brief, 5-8 job stories with a ranked hiring-criteria inventory, a concept viability score per tested idea, and a quantitative segment profile with demand indicators. Enough to brief design, engineering, and marketing from one document.
Common pitfalls
Skipping desk research — you end up rediscovering what competitors already documented three years ago.
Running the survey before the interviews. Your questions won't match the words your customers use, and the data looks clean but means little.
Treating N=30 concept-test feedback as a go/no-go signal. It's a pattern, not a decision.
Redesign an interface
An existing product surface feels dated or confusing, but you don't want to throw everything out. You need evidence for what to keep, what to change, and whether the change paid off.
Method chain
Expected deliverables
A severity-rated backlog from the heuristic pass, a prioritized findings report with video evidence from moderated sessions, and an A/B result with lift, significance, and an implementation recommendation.
Common pitfalls
Running usability tests before heuristic evaluation. Participants burn time on cosmetic violations you could have caught yourself.
Shipping the redesign without A/B testing because 'it's obviously better.' Obvious designs ship regressions, and retention dips a month later tell you too late.
A/B testing a redesign with too little traffic. Week-long tests on a low-volume surface produce inconclusive noise that gets read as approval.
Conversion is dropping
A key flow — signup, checkout, activation — is converting worse than a month or quarter ago. You need to find where and why, fast, before blaming it on acquisition channels.
Method chain
Expected deliverables
A segment-level conversion report, a funnel with step-by-step drop-off and an absolute-user ranking, interview-derived hypotheses for the worst step, and a usability finding list with fix priorities tied to video clips.
Common pitfalls
Jumping to redesign the step before confirming it's a UX problem. Acquisition mix, pricing changes, or outages often masquerade as conversion drops.
Only reading aggregate numbers. The 8% overall drop is a 25% drop for one segment and noise for three others.
Interviewing customers who did convert. You learn why the flow works, not why it's failing for the group you lost.
Enter a new market
You're considering a new geography, segment, or customer type. You need enough evidence to decide whether to commit real resources, and to frame the positioning if you do.
Method chain
Expected deliverables
A 3-10 page market brief with open questions flagged, a competitor matrix with SWOT and a gap map, a contextual persona and artifact inventory from the field, and a segmented survey report with demand indicators by sub-segment.
Common pitfalls
Using a home-market survey to pre-validate. Translations of questions imply different things in different cultures, and the aggregated score hides it.
Skipping field work when the market looks superficially similar to your current one. Unspoken norms are exactly what makes market entry fail.
Treating competitor feature parity as the goal. Feature matrices rarely explain why a market prefers one provider over another.
Measure satisfaction
Leadership needs a tracked UX metric, or a dip in sentiment showed up elsewhere and you need to diagnose it. The goal is a repeatable number plus actionable reasons behind it.
Method chain
Expected deliverables
A benchmarked score per instrument per segment, a coded follow-up theme library tied to the score drivers, and a journey map with a ranked pain-point inventory plus design briefs for the top opportunities.
Common pitfalls
Reporting the score without the follow-up. CSAT of 3.2 is a symptom, not a diagnosis — the number alone produces meetings, not fixes.
Mapping the journey from internal assumptions instead of interview transcripts. The result looks authoritative and points to the wrong pain points.
Choosing NPS by default. It measures loyalty intent, not task satisfaction — for onboarding or a specific feature, CSAT or SUS fit better.
Discover a new feature
The product is stable, usage data is healthy, and the team is running out of clear bets. You need the next feature that earns its build cost rather than filling the roadmap.
Method chain
Expected deliverables
An interview-driven pain-point inventory, a force diagram and hiring-criteria list for the lead job, a proposed IA cluster for the feature, and a concept viability score with kill/pivot/proceed recommendation.
Common pitfalls
Starting with concept testing. Without interview-derived jobs, you test ideas that answer nobody's actual pain and the scores look fine because users are polite.
Stopping at five interviews. Empirical saturation research (Guest 2006, Hennink 2017) puts code saturation around 9-12 interviews — five catches headlines, not patterns.
Letting the most vocal customer shape the roadmap. One passionate user is a sample of one, and card sorting is the cheapest way to see if the idea resonates broadly.
Redesign navigation / IA
Users can't find content. Support sees the same 'where is X' questions on repeat. You need evidence-grounded IA decisions, not another round of internal reshuffling.
Method chain
Expected deliverables
A similarity matrix with a labeled category proposal, a task-level success-rate report from tree testing, a click heatmap and success-zone analysis from first-click tests, and a comparative usability report against the old nav with SUS score and prioritized issue list.
Common pitfalls
Running only card sorting. A clean sort can still produce a tree users can't navigate — validating the tree is where most IA projects fail.
Testing with an over-filtered card set (under 15 items). The analysis is noise; above 80 it exhausts participants — the sweet spot is 30-60.
Skipping the visual-design check. A correct tree rendered with bad labels or weak visual hierarchy fails in first-click testing and you won't know why.
Retention — why users leave
Aggregate DAU looks fine but retention is eroding underneath. You need to see the real curve, find when users disengage, and understand why — before you lose the segment entirely.
Method chain
Expected deliverables
A cohort retention table with curves and behavioral comparisons, a funnel identifying the drop-off stages, a triggers map and timeline from the diary study, and churn-reason stories with segment-specific retention hypotheses ready for A/B testing.
Common pitfalls
Reading aggregate retention only. Healthy acquisition hides cohort decay, and by the time it surfaces you're already six months behind.
Running the diary on a sample that's too small to cover segment variance. Five diaries from one segment tell you one segment's story, not the product's.
Ignoring churn interviews because 'they already left.' Former users are the most candid source you have on what broke the relationship.
Pricing and packaging
You're introducing tiers, repricing, or repackaging an existing feature set. Gut-pricing erodes margin; scientific pricing is more work upfront and better revenue later.
Method chain
Expected deliverables
A current-pricing and feature-importance baseline from survey data, a MaxDiff score list on a -100 to +100 scale with Top-3 and Top-5 reach, a Kano category map with build/invest/defer recommendations, and per-package purchase intent scores.
Common pitfalls
Asking 'how much would you pay for X?' in a survey. Stated willingness-to-pay is notoriously inflated — use MaxDiff trade-offs or conjoint instead.
Treating Kano categories as permanent. Attractive today becomes Performance within a year and Must-be within two. Re-run when competitors close the gap.
Skipping concept testing on the final packages. A mathematically optimal bundle with a confusing name converts worse than a messier bundle users can parse.