By Mehdi HassanWhat is conversion research?
Conversion research is the set of methods used to understand why visitors do or don't take the action a website wants. It combines quantitative data, which shows where and how often people drop off, with qualitative research, which explains what they were thinking when they did.
It's the first and most important stage of conversion rate optimization, because every change you make afterward is only as good as the evidence behind it. Teams that skip research end up redesigning based on opinion, and the loudest opinion in the room is rarely the buyer's. Good research also produces something more valuable than a list of fixes: the buyer's own words, which become your best copy.
What research methods do CRO specialists use?
CRO specialists use six core research methods: analytics and funnel analysis, session recordings and heatmaps, on-site polls, customer surveys, customer interviews and user testing. They're chosen by the question being asked, not run all at once.
Method | Answers the question | Effort | Lightweight version |
|---|---|---|---|
Analytics and funnels | Where do visitors drop off, and which segments? | Low once tracking works | One funnel report by source and device |
Session recordings | What do visitors actually do on the page? | Low to medium | Watch a sample of recordings on one page |
Heatmaps | Where do clicks and attention cluster? | Low | One click map on the weakest page |
On-site polls | What's stopping visitors right now? | Low | One open question on one page |
Customer surveys | Why did buyers choose you, and what nearly stopped them? | Medium | A short email to recent signups |
Customer interviews | How do buyers describe the problem and the decision? | Medium to high | Five calls with recent customers |
User testing | Can people from your audience understand and use the page? | Medium | Three to five unmoderated sessions |
There's an order that works for most teams: analytics first to find the weak page, recordings and heatmaps to see behavior on it, then polls, interviews or user tests to understand why. Our CRO audit process shows how these fit into a full review.
How do you use analytics, recordings and heatmaps?
Use analytics to find where people drop off, then use recordings and heatmaps to see what they do at that point. Neither explains motives, but together they tell you exactly where to aim the qualitative research.
In analytics, map the steps from landing to conversion and compare segments: paid versus organic, mobile versus desktop, new versus returning. A page that converts fine on desktop and badly on mobile is a different problem from one that converts badly everywhere. Then watch session recordings of visitors on that page and segment. Look for hesitation, rage clicks, clicks on things that aren't links, and form fields where people stop.
Heatmaps summarize clicks and scroll depth across many visitors. They're useful for spotting ignored CTAs or sections nobody reaches, but they're easy to over-read. A cold area doesn't always mean a section is bad; it might just not need clicking.
How do you run on-site polls and customer surveys?
Run on-site polls with one short open question shown at the moment of hesitation, and send customer surveys to recent buyers while the decision is fresh. Polls catch visitors who are leaving; surveys catch the ones who stayed.
For polls, ask about the visitor's situation, not your page. "What's stopping you from starting a trial today?" or "What else do you need to know before booking a demo?" work better than "Do you like this page?". Show the poll on the page with the biggest drop, after a reasonable time on page or on exit intent. Our guide to on-site surveys and polls has the questions that reveal the most.
For customer surveys, send a short email to people who signed up or bought recently. Ask what problem made them start looking, what alternatives they considered, what nearly stopped them and how they'd describe the product to a colleague. The answers to the last question are often better than your current headline.
Survey questions that work for SaaS buyers
These open questions work well for recent customers:
1. What were you using or doing before you found us? 2. What made you start looking for a new solution at that point? 3. What other options did you consider? 4. What almost stopped you from signing up? 5. What's the main thing you get from the product now? 6. How would you describe us to a colleague in one sentence?
Keep it to these six or fewer, make every question open rather than multiple choice, and read the answers yourself rather than only looking at a summary.
How do customer interviews and user tests differ?
Customer interviews explore how buyers think about their problem and decision; user tests watch people try to use a specific page or flow. Interviews tell you what to say; user tests tell you whether people understand how you've said it.
Customer interviews
Talk to recent customers and, if you can, prospects who chose someone else. Ask about the situation before they bought, the trigger, the alternatives and the doubts. Record and transcribe the calls so you can quote exact phrases. This is the core of voice of customer research, and a handful of good interviews usually changes the homepage more than any amount of analytics.
User tests
Give people from your target audience a realistic task, such as "find out whether this tool works with your CRM and what it costs", and watch them try. Moderated tests let you ask follow-up questions; unmoderated tests are faster and cheaper. Even a few sessions reliably reveal confusing navigation, unclear copy and broken flows. For checking functionality before launch rather than persuasion, see website QA testing.
What's a lightweight research plan for early-stage teams?
An early-stage team can run useful conversion research in a couple of weeks with one poll, five customer conversations and a review of sales call notes. You don't need large samples; you need repeated patterns from the right people.
Fix tracking and pick the weakest page
Confirm conversions are counted correctly, then find the page with the biggest drop toward your main conversion.
Watch recordings on that page
Note where people hesitate or leave and what they click.
Add a one-question poll
Ask what's stopping them today. Leave it running until you see answers repeat.
Talk to five recent customers
Ask about the trigger, alternatives, doubts and how they'd describe you. Record and transcribe.
Read the last month of sales notes
List the objections and questions prospects raise most often.
Synthesize into hypotheses and copy
Group findings into themes, write a hypothesis for each, and collect the best buyer phrases for the rewrite.
This approach suits sites that can't run statistically valid A/B tests, which is most startups. CRO for low-traffic sites shows how to measure the changes that follow. If you do have the traffic, research gives you the hypotheses worth testing; our A/B testing guide for landing pages takes it from there, and A/B testing mistakes covers how tests built on weak research produce false winners.
How do you turn research into conversion improvements?
Turn research into improvements by grouping findings into themes, writing each as a hypothesis, and ranking them by evidence and impact. Then change the copy first, because most research findings are about what buyers need to hear.
A good hypothesis links evidence to a change and an expected result: "Poll answers and sales notes show buyers don't know if we integrate with their CRM; adding integrations to the hero will increase demo requests from paid search." Score hypotheses with a framework such as the LIFT model, covered in our guide to CRO heuristics, so the strongest go first. Keep the raw quotes; they're the material for the rewrite.
