Conversion research methods are the tools and techniques used to understand why visitors are or are not converting β the essential first step before any A/B test or optimization decision. Without research, CRO becomes guessing. With it, every test is grounded in specific evidence about real buyer behavior. Effective CRO programs combine quantitative methods (what is happening, in numbers) with qualitative methods (why it is happening, in words and behaviors). Teams that master both research types generate higher-quality hypotheses, run fewer failed tests, and compound their learning faster.
Quantitative Research Methods
Quantitative methods produce numerical data about visitor behavior. They tell you what is happening and how often, but not why.
1. Funnel Analysis
What it tells you: Where visitors drop off in the conversion process, expressed as percentages.
Tools: Google Analytics 4 (Exploration β Funnel exploration), Shopify Analytics (funnel report), Mixpanel, Amplitude
How to run it:
- Define your funnel steps: Landing page β Product page β Add to cart β Checkout β Purchase
- Build the funnel in your analytics tool
- Identify the step with the highest absolute volume of drop-off (not just highest percentage)
- Segment by device, traffic source, and visitor type (new vs. returning)
What to do with it: The step with the highest drop-off becomes the priority page for qualitative investigation. If 60% of add-to-cart events result in checkout abandonment, your qualitative research should focus on the cart/checkout experience.
Indian D2C benchmark context:
- Product page β Add to cart: 5β12% (median ~8%)
- Add to cart β Checkout: 35β55%
- Checkout β Purchase: 70β85%
2. Heatmaps
What it tells you: Where visitors click, how far they scroll, and where they focus attention.
Tools: Hotjar, Microsoft Clarity (free), Mouseflow, FullStory
Three heatmap types:
Click heatmap: Shows where users click. Reveals:
- Elements that look clickable but are not (users clicking non-linked images)
- CTAs being missed (low click density on primary button)
- Navigation being used in unexpected ways
Scroll heatmap: Shows how far users scroll. Reveals:
- Whether key content (reviews, trust signals, CTA) is below the scroll depth of most visitors
- Where engagement drops off (the point where most users stop scrolling)
- Mobile vs. desktop scroll depth differences
Attention heatmap (eye-tracking simulation): Combines scroll and time data to estimate where users focus. Most relevant for landing page and product page layout decisions.
3. Click Tracking and Event Analysis
What it tells you: Which specific elements generate clicks, in what sequence, and at what rates.
Tools: GA4 event tracking, Hotjar click analytics, Heap, FullStory
Key events to track for ecommerce:
- Product image gallery clicks (are users engaging with all photos?)
- Review tab clicks (are users reading reviews before adding to cart?)
- Shipping/returns link clicks (are users seeking reassurance pre-purchase?)
- Add to cart button clicks vs. buy now clicks
- Payment method selection patterns at checkout
What to do with it: High click rates on the return policy link (indicating buyers seeking reassurance) suggest making the policy more prominent. Zero clicks on a "Product Details" tab suggest the content is either obvious or irrelevant.
4. Site Search Analysis
What it tells you: What visitors are looking for but cannot easily find through navigation.
Tools: GA4 site search, Shopify search analytics, SearchPie
High-value site search patterns:
- Visitors searching for products that exist but are not easy to navigate to (navigation/category structure issue)
- Visitors searching for non-existent products (demand signals for new products)
- Visitors searching for "return," "refund," "COD" (anxiety signals β they are looking for information your pages do not surface clearly enough)
5. Cohort and Segmentation Analysis
What it tells you: How different buyer groups behave differently.
Tools: GA4 explorations, Shopify analytics, Klaviyo
Valuable segmentations for Indian D2C:
- COD vs. prepaid buyers (very different LTV and behavior patterns)
- Tier 1 city vs. Tier 2/3 city visitors
- First-time vs. returning visitors
- Traffic source cohorts (paid vs. organic vs. email)
- Acquisition month cohorts (do buyers from January have different 90-day LTV than buyers from October?)
Qualitative Research Methods
Qualitative methods produce explanations of behavior β the reasons behind what the numbers show. They tell you why visitors are dropping off, what they understand, what confuses them, and what they need to feel confident enough to buy.
6. Session Recordings
What it tells you: Exactly what individual visitors do on your site β clicks, scrolls, hesitations, and navigation paths β in real time.
Tools: Hotjar, Microsoft Clarity (free), FullStory, LogRocket
How to use session recordings effectively:
Do not watch recordings randomly. Create behavior-based segments:
- Non-purchasers who visited the product page
- Cart abandoners (visited cart but did not start checkout)
- Checkout abandoners (started checkout but did not complete)
- High-time-on-page visitors who did not convert
Watch 20β30 recordings from each segment. Note:
- What did they click on that did not lead them forward?
- Where did they hesitate (cursor pausing, rapid scrolling)?
- What was the last action before leaving?
- What content did they engage with deeply (reviews, FAQs, product images)?
Common findings in Indian D2C session recordings:
- Mobile users trying to tap the delivery date area (expecting it to be interactive)
- Users opening the review section and spending 1β2 minutes reading before leaving (unaddressed concern in reviews)
- Users clicking the back button after reaching the payment page (payment method they want is not visible)
7. On-Site Surveys
What it tells you: Self-reported reasons for behavior, in visitors' own words.
Tools: Hotjar surveys, Typeform, Google Forms (deployed via CustomFit.ai)
Three survey types for ecommerce:
Exit intent survey (non-purchasers):
- Single question: "What stopped you from completing your purchase today?"
- Open text or multiple choice with "Other" option
- Deploy on exit intent for product page and cart page separately
- Collect 50β100 responses before analyzing
Post-purchase survey (new buyers):
- "What almost stopped you from buying today?"
- "Where did you first hear about us?"
- "What was the main reason you chose to buy?"
Abandoned cart recovery survey:
- Via email, 24 hours after abandonment
- "We noticed you didn't complete your order. Can you tell us why?"
- Even a 5β10% response rate yields high-quality qualitative data
Analyzing survey responses:
Cluster responses into themes (shipping cost, payment concern, trust/brand unfamiliarity, product uncertainty, just browsing). The theme with the most responses is your top friction priority.
8. User Testing
What it tells you: How real users interact with your site when given a specific task β in real-time, with thinking-out-loud narration.
Tools: UserTesting, Maze, Lookback, or informal testing via Zoom
Task-based user test format:
- Recruit 5 users who match your target buyer profile
- Give them a specific task: "You are looking for a moisturizer for oily skin under βΉ1,000. Find a product and complete a purchase."
- Ask them to think out loud as they navigate
- Do not guide or help β observe what confuses them
What you discover:
- Whether your navigation and search are intuitive
- Whether your product descriptions answer the buyer's questions
- Whether your trust signals are visible and credible
- Specific copy that creates confusion or misleads
5 user tests typically reveal the most important usability issues. For high-traffic sites, expand to 8β10 users for more diverse coverage.
9. Customer Interviews
What it tells you: Deep motivations, decision-making process, and vocabulary of your buyer β the richest qualitative data source.
Format:
- 20β30 minute phone, video, or WhatsApp call with recent buyers (< 30 days since purchase)
- 5β8 interviews per research cycle
- Open-ended questions: "Walk me through how you decided to buy this product." "What was the moment you decided this was right for you?" "What were you unsure about?"
Especially valuable for:
- High-ticket product pages (understanding the multi-session research journey)
- New product launches (what vocabulary do buyers use? What concerns arise?)
- Post-negative-review analysis (what went wrong, in buyers' words?)
10. Support Ticket Analysis
What it tells you: Common pre-purchase questions, product concerns, and post-purchase dissatisfiers β in buyers' own words, from buyers motivated enough to contact you.
Process:
- Review the last 100 support tickets
- Tag each ticket: pre-purchase question / post-purchase issue / return request / delivery concern
- For pre-purchase questions: each frequently asked question is a friction point not addressed on the relevant page
Common Indian D2C support ticket themes:
- "Is COD available in [city]?" β Add city-level COD indicator to product pages
- "What are the ingredients?" β Ingredient list needs to be more prominent or detailed
- "How long is the return window?" β Return policy needs to be more visible
Combining Quantitative and Qualitative: The Research Sequence
Effective CRO research follows this sequence:
Step 1: Quantitative analysis identifies where the problem is (funnel drop-off analysis, heatmaps)
Step 2: Qualitative research reveals why the problem exists (session recordings, surveys, user tests)
Step 3: Hypothesis formation combines both: "[Specific behavior observed in session recordings] is causing [specific drop-off identified in funnel analysis], so [specific change] should improve [specific metric]"
Step 4: A/B test validates the hypothesis
Step 5: Results return to quantitative analysis to confirm the improvement
This cycle repeats continuously. Each completed experiment generates new quantitative data that informs the next round of qualitative investigation.
Key Takeaways
- Quantitative methods (funnel analysis, heatmaps, click tracking) tell you what is happening; qualitative methods (recordings, surveys, user tests) tell you why
- Start with funnel analysis to find the highest-drop-off stage, then use qualitative research on that specific stage
- 5 user tests reveal ~85% of major usability issues β you do not need large samples for qualitative validity
- On-site exit surveys ("What stopped you from buying?") are the highest-value-per-hour qualitative research tool for most ecommerce brands
- Support ticket analysis reveals pre-purchase friction in buyers' own words β convert each common question into a product page content improvement
- The research cycle (quantitative β qualitative β hypothesis β test β validate) is the foundation of a systematic CRO program
Related reading:
cro
cro
cro