To find out why shoppers aren't buying, ask the people who almost didn't buy, read what your customers write in reviews and support tickets, and watch where visitors stall on the page. Each source shows a different part of the problem. Together they give you objections you can test, not guesses.
This is what CRO research means for an ecommerce store: collecting the reasons behind your numbers before you change anything. Analytics can show you where people drop off. Customer research shows you why.
Practitioners in our database get some of their biggest wins this way. Kasper Dolk reported $763,679.14 extra profit per year from hero bullets written to answer objections customers raised in surveys. Below are the five sources, what to ask or look for in each, and how to turn the answers into tests.
Post-purchase surveys: ask buyers what almost stopped them
A post-purchase survey runs on the thank-you page or in the order confirmation email. Your buyers just made the decision, so they remember what nearly stopped them. They also answer at much higher rates than random site visitors do.
Keep it to one to three questions. Long surveys get skipped, and the first open-text answer is usually the most useful one.
Post-purchase survey questions for ecommerce
- "What almost stopped you from buying today?" This is the most useful question you can ask. The answers become your list of objections.
- "What made you choose us over other options?" These answers give you the reasons to put in headlines and USP bars.
- "What did you hope this product would help with?" This gives you outcome language in the customer's own words.
- "How did you first hear about us?" This is useful for attribution, but it won't give you test ideas by itself.
- "Did you have any questions we didn't answer on the site?" These answers show gaps in your FAQ and product page content.
Use open text for the objection question. Multiple choice is faster to analyze, but you only get back the options you thought of, and the point of the survey is to learn what you didn't know.
The answers turn into tests easily. Oliver Kenyon put three survey-based reassurance lines above a collection page grid, a test recorded at $44,890/month. Kasper Dolk used survey answers to build a "how you'll feel" timeline on the product page, with CVR up 14% (still confirming at the time it was shared).
Surveys can also tell you what to take away. Kasper Dolk's removal of a countdown timer from checkout, a test that drew on survey feedback, recorded CVR +25.61%, RPV +26.18%, est. +€150k/month. For more on setting up surveys, see our page on post-purchase surveys.
Review mining: read what customers already told you
Your product reviews, and your competitors' reviews, are free customer research. Customers write them without being prompted, so they show what people actually care about.
Export reviews into a spreadsheet and tag each one. A simple set of tags is enough:
- Doubts before buying: "I wasn't sure it would fit," "I was skeptical at first."
- Outcomes: what changed for them, in their words.
- Complaints: shipping, sizing, quality and anything else that comes up more than once.
- Comparisons: other brands or products they mention.
Count the tags. The doubts that come up most often are the objections your product page needs to answer near the buy button.
Reading 3-star reviews of competing products is often the fastest way to find out what shoppers in your category are afraid of. Those reviewers bought, were let down in a specific way, and usually explain how.
Oliver Kenyon tested exactly this kind of placement: an objection-handling review block below the CTA on mobile recorded +$63,735/month. Kasper Dolk's outcome headline with a tagline handling the #1 objection at the top of the product page saw a ~10% conversion increase across multiple clients.
Research tells you what to say. The test tells you whether the way you said it works. Browse more results on our reviews page.
Session recordings and heatmaps: watch where shoppers stall
Surveys and reviews tell you what people think. Recordings and heatmaps show what they do, and the two don't always match.
Look for these patterns:
- Repeated clicks on things that aren't links, such as images, badges or USP icons. Shoppers expect more information there.
- Scrolling back up to the price or the variant selector, which usually means they're confused.
- Hesitation on the size chart, shipping line or returns link, which points to a doubt you can answer higher on the page.
- Where attention stops on long pages. Content below that point is close to invisible.
Watch recordings with a question in mind, for example "what do mobile visitors do after adding to cart?" Filtered sets of 20-30 sessions teach more than hours of random viewing.
Dave Diederen has run several tests built on heatmap data. Changing the above-the-fold USPs to the points heatmaps showed buyers engaging with gave +6.23% CVR. Placing a countdown timer where heatmaps showed shoppers were focused gave +20% conversion rate. Turning a collection banner that drew attention into a clickable countdown timer brought +$161K in revenue.
Which tool to use
Several tools record sessions and build heatmaps. Here is how three common options compare, based on vendor pages as of November 2026.
| Tool | Starting price | What stands out | Not a good fit for |
|---|---|---|---|
| Microsoft Clarity | Free (vendor page) | Heatmaps and recordings, with "no limits on traffic" | Stores that want revenue tied to clicks, or surveys and chat in the same tool |
| Lucky Orange | $32/month billed annually for 3,500 sessions (pricing) | Heatmaps, recordings, surveys, live chat and funnels on every plan | High-traffic stores on a tight budget, since price scales with sessions |
| Heatmap.com | $199/month for stores with $0-10M revenue (pricing) | Revenue-based heatmaps and recordings tracked against revenue, plus surveys | Small or new stores, where a free tool gives most of the same insight |
Lucky Orange offers a 7-day free trial, and Heatmap.com offers a 14-day trial, according to their pricing pages. Lucky Orange's lowest plan caps sessions, so a busy store may need a higher plan before the data is complete. Heatmap.com prices by store revenue, not traffic, which makes it expensive for smaller brands. For a fuller comparison, see our heatmap tools roundup.
Support inbox: the questions people ask before buying
Every pre-sale email, chat and DM is an objection your site didn't answer. Ask your support team, or search your helpdesk, for the questions that come before an order.
Group them the same way you tagged reviews. Common groups include shipping times, returns, sizing, ingredients or materials, whether the store is legitimate, and how the product compares to an alternative.
The fix is often to answer the question on the page, near the point where it comes up. Dave Diederen added a trust-focused FAQ section to the homepage covering shipping, returns, payment and legitimacy, which recorded +$8,984/month. Kasper Dolk tested moving the FAQ higher and saw +13.08% RPV.
Support data also tells you who your customers are. Dave Diederen increased product page font size by 4px for an older audience and saw +4.17% RPV, a change you'd only think of if you knew your buyers.
Don't assume more support tools always help. Oliver Kenyon removed a live chat feature that was costing revenue and recorded +$28,000/month. Read the transcripts, then decide whether the answers belong on the page instead. Our FAQ tests page has more examples.
Turning answers into tests
Research is only useful if it changes what you test. Use a simple process:
- Combine your sources in one sheet. List each objection or friction point, with a count from surveys, reviews, recordings and support.
- Rank by frequency and by where it shows up. An objection that appears in three sources and blocks the add-to-cart step comes first.
- Write a hypothesis for each. For example: "Because 30% of survey answers mention sizing doubts, adding a fit guide next to the size selector will raise add-to-cart rate." Our hypothesis guide shows the format.
- Use customer words in the variant. Copy phrases straight from reviews and survey answers.
- Prioritize and test. Score ideas by impact and effort with a framework from our prioritization guide, then check you have enough traffic with the sample size calculator.
Oliver Kenyon's four customer-led USP icons beneath the cart checkout button show the full loop: research found the reasons customers buy, and the test put them where the decision happens. It recorded +$21,516/month.
Expect some research-backed ideas to lose. That's still useful, because a loss rules out a wrong belief about your customers. Dave Diederen, for example, saw -3.84% RPV (loss) after adding small customer profile icons next to the star rating.
@oliverkenyon@oliverkenyon@oliverkenyon
@DaveDiederen@DaveDiederen
@kasper_dolk