Ecommerce Conversion Research: How to Find Out Why Shoppers Don't Buy

Conversion research finds why shoppers don't buy by combining methods. Analytics shows where visitors drop off and which segments struggle. Session recordings and usability tests show how they struggle. Surveys, interviews, reviews and support tickets explain why in customers' words. Heuristic and accessibility reviews catch known issues, and technical checks find speed, error and device problems. Triangulate: a finding supported by two or more methods is strong evidence. Turn findings into hypotheses, fix clear problems and test uncertain ones.

Tests based on opinions or competitor copying win rarely and teach little. Tests based on research start from observed problems, so they're more likely to address real doubts, and even their losses narrow down what's going on. Research also finds problems that don't need testing at all: bugs, errors and missing information that should simply be fixed.

This article describes research methods and how to combine them. For the full audit process, see ecommerce CRO audit; for turning findings into tests, see experiment prioritization.

Analytics shows where problems occur. Start with the funnel from landing through product view, add to cart, checkout and purchase, segmented by device, traffic source, new vs returning and market. Look for steps where one segment drops more than others. Check landing pages with high exits, search terms with zero results, and product pages with views but few add to carts.

Analytics doesn't explain why. Its job is to direct qualitative research to the right places. See ecommerce conversion funnel and search analytics.

Recordings show individual sessions: where people scroll, hesitate, rage-click, hit errors or leave. They're most useful when filtered to a question, such as mobile sessions that reached checkout and left at the delivery step. Heatmaps aggregate clicks and scroll depth, showing whether key content is seen. Configure tools to mask personal data. See ecommerce heatmaps.

Usability tests watch real people attempt tasks: find a product for a given need, choose a size, check delivery, complete checkout. Moderated sessions allow follow-up questions; unmoderated sessions scale more cheaply. Small rounds with around five participants per key segment often surface the most common problems, and several rounds over time are more useful than one large study.

Recruit people who resemble your shoppers, use realistic tasks, and avoid leading questions. Record what participants do, not just what they say. See usability testing and user research methods.

Surveys capture reasons at scale. On-site surveys ask a single open question at a relevant moment: on product pages ("Is anything unclear about this product?"), on cart exit ("What's stopping you from completing your order?"). Post-purchase surveys ask what nearly stopped the purchase, what alternatives were considered and how the customer heard about the store. Keep surveys short, open-ended at first, then add structured options once themes are known.

Customer interviews go deeper: how shoppers chose, what they compared, what worried them, what they wish had been clearer. They're especially valuable for high-consideration products and B2B buying, where decisions involve research and several people. Interview recent customers and, where possible, people who considered buying but didn't.

Customer service tickets, live chat logs, product reviews and return reasons are research that already exists. Tag them by topic (delivery, sizing, product information, payment, returns) and count by product and period. Frequent pre-purchase questions show missing information; frequent return reasons show expectation gaps that product pages could close.

An expert review against usability principles and ecommerce best practices finds common problems quickly: unclear navigation, missing information, weak error messages, inconsistent patterns. An accessibility review against WCAG finds barriers that affect people with disabilities and often everyone else too. Both rely on judgement, so treat findings as hypotheses unless they're clear failures. See ecommerce accessibility checklist.

Technical problems are among the easiest wins: slow pages on mobile, JavaScript errors on specific browsers, broken layouts on certain devices, failing payment methods, tracking gaps. Check page speed on real devices, error logs by browser, and conversion by browser and device for outliers. See page speed and conversion.

Each method has blind spots. Analytics shows what but not why; surveys show what people say, which may differ from what they do; usability tests are small; heuristic reviews are opinion. Combine them. A finding supported by two or more methods (analytics shows drop-off at delivery, surveys mention delivery costs, recordings show shoppers searching for delivery information) is strong evidence and should rank high in the backlog.

Research involves real people and their data. Get informed consent for usability tests and interviews, mask personal data in recordings, keep survey data secure, follow your privacy notice and applicable law (which varies by jurisdiction), and don't record more than you need. See ecommerce privacy and customer data.

Plan research like a project, with questions, methods, participants and outputs. A plan keeps research focused and makes it repeatable each quarter.

Research produces a lot of notes. Synthesize them into findings: each finding states the problem, where it occurs, the evidence from each method, its likely impact and a suggested next step (fix, test or research further). Group findings by theme and by template. A synthesis workshop with design, development and marketing helps turn findings into shared understanding and a backlog. See experiment prioritization.

Small teams can still do meaningful research. Analytics and support tickets cost nothing extra. A single open question on the thank-you page takes minutes to set up. Five remote usability sessions with recent customers can be run in a week. Heuristic reviews against published ecommerce guidelines are quick. The key is to ask focused questions and combine at least two sources before acting.

Conversion research combines analytics, recordings, usability tests, surveys, interviews, support data, expert reviews and technical checks. Triangulate findings, fix clear problems and turn the rest into evidence-backed hypotheses. Related: CRO testing roadmap and customer analytics.