E-commerce conversion rate optimisation: measuring and fixing the funnel
How to grow revenue without growing traffic. Measuring funnel steps, finding drop-off points, prioritising tests, and the 12 improvements that actually work.
Most e-commerce teams think about growth with one lever: more traffic. Yet growing revenue on the same number of visitors is usually shorter and cheaper. Lifting conversion from 1% to 1.5% produces the same result as increasing traffic by 50% — and requires no ad budget.
This article covers how to measure the funnel, find the drop-off points, and prioritise the fixes.
Summary
- Conversion rate is not one number; it is the step-to-step rates through the funnel.
- Pouring ad budget into a leaking funnel magnifies the loss — funnel first, traffic second.
- The three most common causes of cart abandonment are surprise shipping, forced signup and long forms.
- Without enough traffic, qualitative methods and obvious-defect fixes beat A/B testing.
- Speed is not a CRO tactic but a precondition; tests run before it is fixed are noisy.
Conversion rate is not a single number
"Our conversion rate is 1.2%" tells you nothing on its own. What matters is the transition rate at each funnel step — because only then can you see where the loss is.
| Funnel step | What it measures | Typical cause of loss |
|---|---|---|
| Visit → product view | Category and search quality | Weak navigation, poor search |
| Product → add to cart | Persuasiveness of the product page | Missing images, unclear stock, low trust |
| Cart → checkout start | Cart transparency | Surprise shipping, hunting for coupons |
| Checkout → order placed | Form and payment flow | Forced signup, long forms, error messages |
| Order → repeat order | Delivery and aftercare | No communication, difficult returns |
The chart below shows an example funnel. The numbers are a scenario model, not a measurement; you need to derive your own from your analytics.
Most people looking at this table focus on the last step. Yet the largest absolute loss is at the second: of 4,200 people, only 980 add to cart. Look at the number of people lost, not the percentage — that is where the effort belongs.
Where to start: three sources
1. Quantitative: analytics
Set up the funnel report and view it split by device. If mobile converts at a third of desktop, the problem is not your product but the mobile experience. A traffic-source split is equally essential: paid and organic visitors do not travel the same funnel.
2. Qualitative: session recordings and user testing
Watching five people go through checkout usually tells you more than a month of A/B testing. What you are looking for is not the average but the moment of friction: the user who goes back, fills the same field twice, or stalls on the page.
3. Direct: exit survey
One question asked of the abandoning visitor — "what stopped you completing your order today?" — turns your list of guesses into a list of facts.
Prioritising the fixes
To decide what to do with limited time, a simple score is enough: impact × confidence ÷ effort. The chart below ranks common levers on that logic.
The last row is deliberate: button-colour tests are the most popular and lowest-yield work in CRO. Testing button colours before removing a surprise shipping charge is painting the oars on a leaking boat.
Twelve improvements that actually work
Product page
- Show shipping cost and delivery time next to the price. A charge that appears at checkout is the leading cause of abandonment.
- State stock status accurately. Not "only 3 left" — the truth. False urgency destroys trust.
- At least five images, ideally one showing scale. Real-world size is the most asked question.
- Put the returns policy on the product page. One sentence, not a link to another page.
Cart
- Show subtotal, shipping and total separately. No surprises.
- Shrink the coupon field. A large coupon box sends the couponless visitor off to hunt for one, and they do not come back.
- Keep "continue shopping" available after adding to cart. Do not force the checkout path.
Checkout
- Allow guest checkout. Account creation can be offered after the order.
- Cut form fields to what is genuinely needed. Every field is an abandonment risk.
- Put error messages next to the field, in plain language. Not "invalid input" but "Card number must be 16 digits".
- Use autocomplete on address fields. The biggest time sink on mobile.
General
- Measure mobile speed and tie it to a budget. We covered the method in how to build a fast website.
The compounding effect of small gains
A 10% improvement at one step looks trivial; the same gain across four steps changes revenue noticeably.
A 10% gain at four steps lifts orders by 46% — with no single heroic change. That is the logic of CRO: not a big idea, but accumulated small corrections.
Thinking about traffic and conversion together
CRO is not an alternative to ad management but a precondition for it. Once the funnel converts, the same ad budget produces more orders. On Ambalaj Cini we ran the e-commerce build alongside Google Ads and Meta campaigns; the 30% growth the company achieved came from both working at once.
We often see the reverse too: ad spend rising, revenue flat. In that picture the problem is almost never the ads — it is the page the ads point to.
Measurement discipline
| Rule | Why |
|---|---|
| At least 2 weeks of baseline before changing anything | Without a reference point, any claim of improvement is empty |
| One major change at a time | With two changes you cannot tell which one worked |
| Report mobile and desktop separately | The average hides the mobile problem |
| Account for seasonality | Compare a campaign period against the same period last year |
| Track profit per order, not revenue | Free shipping can raise conversion and lower profit |
That last row matters most: not every change that raises conversion is profitable. Judge on contribution per order, not on revenue.
Conclusion
Conversion optimisation is not about finding a magic button colour; it is about measuring where the funnel leaks and closing that hole. The order is fixed: measure, find the largest absolute loss, make the low-effort fixes, measure again. Once that loop is running, growth stops depending on ad budget.
If you would like us to build your store's funnel report together, write via the contact page. For the technical side see our web application and web design services, and for the technology choice read Next.js or WordPress.
Frequently asked questions
- What is a good conversion rate?
- It varies so much by sector, traffic source and basket size that a single target number is misleading. The comparison that matters is not an external benchmark but your own history: what happened versus last quarter on the same traffic source. Instead of fixating on the absolute number, look at which funnel step you are losing people on.
- Should I increase traffic first or fix conversion?
- Almost always conversion first. Lifting conversion from 1% to 1.5% produces the same revenue as increasing traffic by 50%, at far lower cost. And pouring ad budget into a leaking funnel only makes the loss bigger.
- How much traffic do I need for an A/B test?
- Roughly, with a low baseline conversion rate you need a few thousand conversions per test, which takes weeks for most small and mid-sized stores. Without that volume, qualitative methods (session recordings, user testing, exit surveys) and fixing obvious defects produce results faster than A/B testing.
- How do I reduce cart abandonment?
- The three most common causes are surprise shipping costs appearing at checkout, forced account creation, and long forms. All three are design decisions: show shipping cost on the product page, allow guest checkout, and cut form fields down to what is genuinely required.
- Does site speed really affect conversion?
- Yes, and most of all on mobile. A slow page does not only lose the impatient visitor; a delay at the payment step turns a completed order into an abandoned one. Speed is a prerequisite for conversion work — tests run before it is fixed produce noisy results.