Framework · Measurement · 2 min read
Qual and quant balance
Quant shows where and how much; qual explains why. Use quant to find and verify problems, and qual to understand and solve them.
- 1Find
Quant shows where users struggle and how many are affected.
"Where is the drop-off?" - 2Understand
Qual shows why it happens and what the problem really is.
"Why do they stop here?" - 3Verify
Quant again: a benchmark shows whether the fix worked.
"Did success go up?"
Repeat: the next benchmark shows you where to look next.
Use it when
- Analytics show a drop-off or a slow step, and the team starts guessing at the cause.
- A stakeholder dismisses research findings as "just five users".
- You need to prove that a redesign worked, not only that users liked it.
How to use it
- Find the problem with quant. Use analytics (where allowed), support tickets or a benchmark to see where users struggle and how many are affected. Output: a ranked list of problem areas with numbers.
- Understand it with qual. Run usability tests, interviews or Contextual inquiry on the top problem to learn what the problem actually is and why it happens. Output: the cause, with observed examples.
- Solve it. Design the fix from what you learned, not from the number alone. Output: a design change aimed at the cause.
- Verify with quant. Re-measure with the same method, for example a UX benchmarking round, to see if the number moved. Output: a before/after result.
- Explain the result with qual. If it worked, short sessions tell you why; if not, they show what you missed. Output: a short note on why the change worked or didn't.
Worked example
On an internal operations platform, the questions each method answers:
| Quant tells you | Qual tells you |
|---|---|
| What and how much | Why and how |
| Where the problem is | What the problem is |
| Whether a change worked | Why it worked (or didn't) |
Support tickets show that many tickets per month are about one customs screen (quant). Two hours of shadowing operators show they copy codes from a spreadsheet because the lookup is hidden (qual). After the fix, a benchmark shows task success up (quant). For internal tools, a few hours of shadowing often teaches more than months of click data.
Watch out for
- Fixing from the dashboard. A number shows where, not why. Talk to users before you design the fix.
- Counting qual findings as statistics. "4 of 5 users" shows a problem exists, not how common it is. Use quant for size.
- Only qual at the end. "Users liked it" doesn't show improvement. Plan a quant measure before the change.
- Waiting for analytics. If tracking is not allowed, use tests, tickets, surveys and observation instead.
Template
Copy this into your research plan for each problem you investigate.
Problem area:
Quant: where / how much? Source: Number:
Qual: why / how? Method: Finding:
Fix:
Verify (quant): metric before -> after
Explain (qual): why it worked / didn't