Conversion

Conversion analysis

On this page

What is conversion analysis?
How to set goals and event tracking before you analyse
Quantitative analysis: what does the data show?
Qualitative analysis: why do users behave the way they do?
The analysis process step by step
Problems conversion analysis usually uncovers
Which tools do you need for conversion analysis?

Conversion analysis is the process of understanding user behaviour: what people do on the website, where they get stuck, why they leave, and what distinguishes those who convert from those who do not.

Without analysis, conversion rate optimisation (CRO) is guesswork. With analysis, the work is based on actual behaviour and evidence.

A thorough conversion analysis combines quantitative data (what happens, where it happens, how often) with qualitative data (why it happens, what users are thinking). Together, they show where the problems are and what needs to be fixed.

What is conversion analysis?

Definition

Conversion analysis is a systematic evaluation of user behaviour on the website in order to identify:

Where users drop out of the conversion funnel
Why they drop out (friction, confusion, lack of trust)
What distinguishes converting users from non-converting users
Opportunities for improvement

Two types of analysis

Quantitative analysis (what and where): Data from analytics tools that shows figures and patterns.

Example

1,000 visitors on a product page
200 add to basket (20 per cent of visitors)
30 complete a purchase (15 per cent of those who added to basket)

Qualitative analysis (why): An understanding of users’ thoughts and motivation.

Example

Session recordings show people hovering over “Add to basket” but not clicking
A user test reveals: “I’m not sure about the delivery cost”
Survey response: “The form was too long”

Both are needed: quantitative analysis shows where the problem is,
and qualitative analysis explains why it is a problem.

How to set goals and event tracking before you analyse

Define goals

Before any analysis can be done, clear goals must be defined.

What counts as a conversion?

E-commerce

Primary goal: Purchase
Secondary goals: Add to basket, begin checkout

B2B

Primary goal: Book a demo, fill in a contact form
Secondary goals: Download a white paper, subscribe to a newsletter

Media

Primary goal: Subscribe
Secondary goals: Register an account, read X articles

Set up event tracking

Google Analytics 4: Configure events for every important step in the funnel.

E-commerce example

view_item (product page viewed)
add_to_cart (add to basket)
begin_checkout (begin checkout)
add_payment_info (add payment details)
purchase (purchase completed)

B2B example

page_view (landing page)
cta_click (click on “Book a demo”)
form_start (starts filling in the form)
form_submit (submits the form)

Why it matters: Without correct tracking, it is impossible to know where users drop out.

Historical data

If goal tracking is already in place, historical data can be analysed.

These are the things to analyse

Conversion rate over time (trends)
Changes after launches
Differences by traffic source
Differences by device (desktop vs mobile)

If there is no history: Start collecting data now. After 2-4 weeks there is enough for an initial analysis.

Quantitative analysis: what does the data show?

Funnel analysis

Visualise the conversion funnel and see where most users drop out.

Example (e-commerce)

Step 1: Product page
1,000 visitors

Step 2: Add to basket
200 visitors (20 per cent continue, 80 per cent drop off)

Step 3: Begin checkout
150 visitors (75 per cent continue, 25 per cent drop off)

Step 4: Payment details
100 visitors (67 per cent continue, 33 per cent drop off)

Step 5: Purchase
30 visitors (30 per cent continue, 70 per cent drop off)

Total conversion rate: 3 per cent
Analysis: The biggest leaks:

Product page → Basket (80 per cent drop off) ← Biggest problem
Payment → Purchase (70 per cent drop off) ← Second biggest

Action: Focus on the product pages first. Why do so few add to basket?

Behavioural analysis: what sets converting visitors apart?

The questions answered

How do people who convert behave on the website?

Analysis in Google Analytics

Average number of pages visited before a purchase
Time on site
Returning vs new visitors

Example

Converting visitors: 6 pages, 8 minutes on site, 60 per cent returning
Non-converting: 2 pages, 1 minute on site, 90 per cent new

Insight: Converting visitors do more research. New visitors may not be ready to buy straight away (they need remarketing).

Do people who convert share common behaviours?

Example

80 per cent of converting visitors read the “How it works” page
70 per cent read customer reviews
50 per cent read the FAQ page

Insight: These pages may contribute to visitors’ trust, but the correlation does not show the cause. Test making them easier to find.

How many visitors who start a conversion complete it?

If 200 start but only 30 complete (15 per cent), there is a large leak in the conversion flow.

Compare

Which sources have the highest completion rate?
Which devices (desktop/mobile) complete more often?

Example

Desktop: 25 per cent complete
Mobile: 8 per cent complete

Insight: The mobile checkout has serious problems.

Do drop-offs happen at a specific point?

Analysis: In the funnel analysis, identify the exact step where most people leave.

Example (checkout)

Step 1 (delivery details): 10 per cent drop-off
Step 2 (payment details): 40 per cent drop-off ← The problem is here
Step 3 (confirmation): 5 per cent drop-off

Insight: Something in the payment step is putting people off.

Traffic sources: where do converting visitors come from?

Analysis by channel

Google Analytics 4 → Reports → Traffic acquisition: Compare key events and conversion rate by source.

Example

Source	Visitors	Conversions	Conversion rate
Organic search	5,000	150	3 per cent
Direct	2,000	80	4 per cent
Paid search	1,500	60	4 per cent
Social media	1,000	10	1 per cent
Email	500	25	5 per cent

Insights

Email traffic converts best (5 per cent): existing customers and subscribers
Social media converts poorly (1 per cent): the wrong audience or too early in the journey
Organic and paid search convert well (3 to 4 per cent): high purchase intent

Landing pages

Which pages do converting visitors start on?

Analysis: Google Analytics 4 → Reports → Landing page. Compare key events by page.

Example

40 per cent start on a specific product page (a bestseller)
30 per cent start on a campaign landing page
20 per cent start on the home page
10 per cent start on the blog

Insight: Some products and landing pages work much better as entry points.

Search behaviour: what are converting visitors searching for?

Search Console shows clicks, impressions, click-through rate and average position, but not conversions. Conversions per search term come from the search terms report in Google Ads. For organic traffic, the link can only be made indirectly: search queries per landing page in Search Console are compared with the same pages’ conversions in your web analytics.

Analysis in Google Ads and your web analytics

Compare search terms

Which search terms lead to the highest conversion?
Generic vs specific searches

Example

Generic search term: “shoes”

10,000 impressions, 500 clicks, 5 conversions (1 per cent conversion rate)

Specific search term: “nike air max 90 women’s size 38”

100 impressions, 50 clicks, 5 conversions (10 per cent conversion rate)

Insight: Specific searches convert much better (the user knows exactly what they want).
Action: Optimise for specific long-tail search terms, not just generic ones.

Brand vs non-brand

Brand searches: The company’s name

Usually convert noticeably better (people are searching specifically for the company)

Non-brand: Generic product searches

Lower conversion (people are comparing alternatives)

Insight: Brand traffic converts well but is limited in volume. Non-brand requires more persuasion.

Qualitative analysis: why do users behave the way they do?

Heatmaps

Show where users click, how far they scroll and where they move the mouse.

Tools: Contentsquare (formerly Hotjar), Microsoft Clarity, Crazy Egg

What to look for

Click heatmaps

Do people click on elements that are not clickable? (confusion)
Are important calls to action (CTAs) being ignored? (poor placement or design)

Scroll maps

How far do users scroll?
Do they see important information (CTAs, trust signals) or leave before then?

Move maps

Where do people hover but not click? (hesitation)

Example: A heatmap shows that 70 per cent of visitors never scroll down to the “Add to basket” button on the product page (the button is too far down).
Action: Move the CTA higher up (above the fold).

Session recordings

See exactly what individual users do on the website.

Tools: Contentsquare (formerly Hotjar), Clarity, FullStory

What you can see

Every click, scroll and form interaction
Repeated clicks (the user clicks the same spot several times in frustration)
Confused movement patterns (jumping back and forth between pages)

Example

The recording shows: A user fills in the form, reaches the “Phone number” field, pauses for a long time, then leaves without submitting.
Hypothesis: “Phone number” feels too personal, or people are unsure why it is needed.
Action: Make the field optional or add explanatory text (“We will only call if there is a delivery problem”).

User testing

Ask real users to test the website while they think aloud.

Platforms: UserTesting, Lookback, Maze

Process

Recruit users (ideally from the target audience)
Give them a task (“Find and buy product X”)
Users carry out the task while putting their thoughts into words
Observe and document problems

Example

Task: “Buy a product”

Users say

“I can’t find the delivery cost… I want to know that before I buy”
“The form is so long… Do I really have to fill in everything?”
“Is this site secure? I don’t see any reviews”

Insights

The delivery cost should be shown early
The form feels too long
Trust signals are missing

Actions

Show the delivery cost on the product page
Shorten the form
Add a Trustpilot widget

Surveys

Ask users directly why they acted (or did not act).

Tools: Contentsquare (formerly Hotjar), Typeform, Google Forms

Types

Exit-intent survey: Triggered when a user tries to leave.
Question: “What stopped you from buying today?”

Answer options

The price is too high
I didn’t find what I was looking for
Unsure about product quality
I need to think about it
Technical problems

Post-purchase survey: After a successful conversion.
Question: “What nearly stopped you from buying?”

The answers can reveal friction that almost prevented the purchase.

On-page survey: Ask directly on a specific page.
Example (checkout page): “Is anything unclear or confusing on this page?”

Competitor comparison

Analyse competitors’ websites to see what they do better (or worse).

What is compared

Checkout flow (how many steps, how simple)
Trust signals (reviews, guarantees)
CTA design and placement
Mobile experience

The aim is not to copy them, but to identify the industry standard and find ways to stand out.

The analysis process step by step

Step 1: Set up tracking

Before the analysis

Define goals and conversions
Implement event tracking (Google Analytics 4)
Install tools for heatmaps and recordings: Contentsquare (formerly Hotjar) or Clarity
Collect data for 2-4 weeks (at least)

Step 2: Quantitative analysis

What is done

Funnel analysis

Identify where most people drop out
Compare completion rates per step

Segment analysis

Compare converting vs non-converting visitors
Compare sources (organic, paid, direct)
Compare devices (desktop, mobile, tablet)

Behavioural analysis

Pages visited before conversion
Time on site
Returning vs new

Result: A list of quantitative insights (where the problem is, how big it is).

Step 3: Qualitative analysis

What is done

Heatmaps

Review click, scroll and move maps for critical pages

Session recordings

Watch 50-100 sessions (prioritise non-converting visitors and those who dropped out at checkout)

User testing

Test with 5-10 users
Identify common problems

Surveys

Exit-intent and post-purchase surveys
Collect 50+ responses

Result: A list of qualitative insights (why the problems exist).

Step 4: Synthesise insights

Combine quantitative and qualitative

Example

Quantitative: 70 per cent drop off at the payment step in the checkout.
Qualitative: Session recordings show people pausing for a long time at the “Create an account” option. A user test reveals: “I don’t want to create an account just to shop.”
Synthesis: Mandatory account registration in the checkout may be contributing to the drop-offs. Test guest checkout to see how much of the drop-off account registration explains.

Step 5: Prioritise problems

Not every problem can be fixed at once.

Prioritisation based on

Impact: How many users are affected?
Severity: How much does it affect conversion?
Effort: How hard is it to fix?

Impact vs effort matrix: Focus on “high impact, low effort” problems first.

Step 6: Document and report

Create an analysis report

Summary

Current conversion rate
The biggest problems identified
Expected impact of the actions

Detailed findings

Quantitative data (figures, charts)
Qualitative data (user quotes, heatmap screenshots)
Recommendations

Next steps

Link to the action plan
A prioritised list of changes to test

Problems conversion analysis usually uncovers

Trust problems

Symptoms

High drop-off on product pages or at checkout
Session recordings: People hover over “Buy” but do not click

Causes

No customer reviews
No payment logos
No visible returns policy or guarantee

Action: Add trust signals.

Form too long or complex

Symptoms

High start rate, low completion rate on forms
Session recordings: People start filling in the form but stop halfway

Causes

Too many fields
Asks for information that is not needed
Unclear labels

Action: Shorten the form, split it into steps, and explain why the information is needed.

Unclear calls to action

Symptoms

Low click-through rate on CTA buttons
Heatmaps: People click around the button but not on it

Causes

The CTA is not visible (poor placement or colour)
Unclear wording (“Continue” instead of “Buy now”)
Several competing CTAs (confusion)

Action: Make the CTA clearer (wording, colour, placement) and focus on one primary CTA.

Slow loading (especially on mobile)

Symptoms

Poor Core Web Vitals in field data and a large gap between clicks and sessions
A high bounce rate, that is, many sessions that do not count as engaged
Mobile conversion much lower than desktop

Causes

Large unoptimised images
Too much JavaScript
A slow server

Action: Optimise Core Web Vitals, especially LCP (Largest Contentful Paint).

Missing information

Symptoms

User tests: “I don’t know X”
Surveys: “Couldn’t find information about Y”

Causes

Delivery cost not visible until checkout
Delivery time unclear
Returns policy hidden

Action: Make important information clear early on (before people make up their minds).

Which tools do you need for conversion analysis?

Analytics

Google Analytics 4

Funnel analysis
Segmentation
Event tracking
E-commerce tracking

Use: Quantitative analysis, identifying where problems are.

Heatmaps & session recordings

Contentsquare (formerly Hotjar)

Heatmaps (click, scroll, move)
Session recordings
Surveys

Microsoft Clarity (free)

Heatmaps
Session recordings
Repeated clicks and clicks with no effect

Use: Qualitative analysis, understanding why problems exist.

User testing

UserTesting

Recruit testers from the target audience
See and hear users navigate the site

Maze

Remote user testing
Prototype testing

Use: Deep qualitative insight from real users.

A/B tests (the next step)

On a modern website, an external testing tool on its own is rarely enough. Instead, A/B tests are usually built into the systems you already have, and the data is extracted for a separate analysis.

Tests are run after the analysis to try out the solutions.

A/B tests are not an analysis tool as such, but they are an important next step.

What to do after the conversion analysis

Conversion analysis is not the end point. It is the beginning.

Process

1. Conversion analysis

Identify where and why problems exist
Document the findings

2. Action plan

Prioritise problems
Plan solutions
Implement

3. Conversion testing

A/B test changes
Validate that the solutions work

4. Measure results

Before/after comparison
Document what worked

5. Iterate

Continuous analysis
New data → New insights → New actions

CRO is a recurring cycle rather than a one-off project.

Contact us about conversion analysis

Carnaby helps companies carry out in-depth conversion analysis, from funnel analysis and behavioural data to session recordings and user testing, to identify exactly where and why users drop off.

Want to know where you stand today?

Call us, or fill in the form and we will get back to you.