Qualitative vs Quantitative Research

Quantitative data tells you WHERE and HOW MANY; qualitative research tells you WHY. You need both — the numbers locate the problem, observation explains it.

Where it comes from

The distinction predates the web — it comes from the social sciences, where researchers have long separated counting things (quantitative) from understanding them (qualitative). User-experience research inherited both traditions: the Nielsen Norman Group has spent decades arguing that the two are complementary, not rival, and that the hard part is knowing which question each one answers.

Why it matters for your website

Two kinds of evidence sit behind every good design decision, and they answer different questions. Quantitative data — analytics, conversion rates, bounce, time on task — tells you WHAT is happening and WHERE: which page leaks visitors, which step in the funnel they abandon, how many never scroll to the offer. It is excellent at locating a problem and almost silent on explaining it. Qualitative research — session recordings, heatmaps, usability testing, user interviews — tells you WHY: you watch someone scroll past the form, hesitate at a field they didn't expect, or click something that looks like a button but isn't. The mistake teams make is treating these as alternatives; they are a sequence. The numbers tell you where to look, then observation tells you what to fix. A chart that says "70% abandon checkout" is a question, not an answer — and redesigning on the strength of the number alone is how a month of work moves the metric by a rounding error. Kweri's whole method follows this order: the audit and your analytics locate the where, each finding's hypothesis proposes the why, and its recommended test confirms it on your own traffic.

The practical trap is that quantitative data is cheap, abundant, and feels authoritative, so teams over-rely on it. A dashboard full of numbers can tell you a page is failing without ever telling you why — and a precise measurement of the wrong thing is still the wrong thing.

Qualitative work is the opposite: slower, smaller-sample, and easy to dismiss as anecdote, yet it's where the actual fix usually hides. Watching five people use a page almost always surfaces a problem no chart would have named. The discipline is to let the numbers point you at where to watch, so the qualitative time is spent on the pages that matter.

Wrong vs right

Wrong

A team sees checkout conversion fall and spends a sprint reordering fields, rewriting copy, and adding a progress bar — all reasonable guesses drawn from the number alone. Conversion barely moves and they can't say which change, if any, helped.

Right

They start from the same number, then watch ten recordings of people who dropped at the payment step. Eight pause at an unexpected "billing address must match card" error that fires too aggressively. One copy fix recovers most of the lost conversions — found in an afternoon, not a sprint.

Wrong

A landing page has a high bounce rate, so the team assumes the headline is weak and A/B tests four new ones. None wins.

Right

A scroll map shows 80% of visitors never reach the part of the page the headline was meant to set up — the problem is page length and load order, not the words. The number said "bounce"; only the qualitative view said where.

Understanding Qualitative vs Quantitative Research

Quantitative and qualitative research are not competing philosophies; they are two stages of the same loop. Quantitative methods — analytics, funnels, conversion rates, time-on-task, rage-click counts — are measurement at scale. They are trustworthy precisely because they aggregate thousands of sessions into a number, and that same aggregation is why they can't explain themselves: a conversion rate has no story attached.

Qualitative methods supply the story. Session recordings and heatmaps show real behaviour on your live site; usability testing and user interviews probe intent and comprehension directly. The cost is sample size and effort, which is exactly why you don't start here — you start with the numbers, find where the experience leaks, and then spend your limited qualitative attention on that specific page or step.

Run in that order, the two compound. The quantitative signal is a hypothesis generator ("something is wrong at step two"); the qualitative review is the diagnosis ("the dropdown doesn't open on mobile"); and a follow-up measurement confirms the fix worked. Skip the qualitative middle and you are optimising by guesswork; skip the quantitative bookends and you are reacting to anecdote. The teams that improve fastest treat them as a single instrument with two lenses.

How Kweri checks it

Kweri is itself mostly a quantitative-and-heuristic instrument: it measures what an automated check can measure — performance, accessibility, technical SEO, and a structured read of the page — and it detects whether you have a qualitative tool (session recording / heatmaps) installed so the other half of the loop is available to you. What it deliberately doesn't do is pretend to do the qualitative work for you: it can't watch your recordings, run a usability test, or interview your users. That's why every finding is framed as a hypothesis with a recommended test — the audit locates the where and proposes the likely why; confirming it is qualitative work only you can do on your own traffic.

FAQ

Which should I do first, qualitative or quantitative?

Quantitative first, almost always. The numbers are cheap to gather and tell you where to look; they turn a vague worry ("the site underperforms") into a specific question ("why do people abandon at step two?"). You then spend your scarcer qualitative time only on the pages the data flagged.

Isn't qualitative research just anecdote?

A single recording is anecdote. But patterns across even five to ten sessions are reliable evidence — usability research has shown for decades that a handful of observed users surfaces the large majority of serious problems. The rigour comes from watching with a specific question, not from large samples.

Can analytics alone tell me what to fix?

It can tell you what to investigate, rarely what to fix. A number locates a problem but doesn't explain it, and the same drop-off can have several different causes that demand opposite fixes. Acting on the number alone is how teams burn a sprint moving a metric by a rounding error.

What's the minimum qualitative setup for a small team?

A free session-recording tool such as Microsoft Clarity, reviewed for thirty focused minutes against whatever your analytics flagged that week. That single habit — numbers point, recordings explain — gives a small team most of the value of a formal research practice at almost no cost.

Related

Attribution & sources

Identified by User-research practitioners; Nielsen Norman Group (methodological grounding). Catalogued from Nielsen Norman Group — "Quantitative vs. Qualitative Usability Testing".

The qualitative/quantitative distinction originates in the social sciences and has no single inventor; its application to UX research is most consistently articulated by the Nielsen Norman Group.

Read the primary source →

See Qualitative vs Quantitative Research on your own site

Run a free Kweri audit — a plain-English review of your site’s speed, accessibility, SEO and design, ranked by what to fix first. No login, no jargon.

Run a free audit →