Conversion Rate Optimization: A 7-Step Process That Does Not Rely on Tricks

Conversion rate optimisation, or CRO, is the practice of getting more of your existing visitors to take a desired action, such as buying, booking or enquiring, by understanding why they do not and testing improvements. It is not about tricks or pressure. It is about removing friction, answering doubts and making the next step clear, then proving with evidence that your changes helped.

Traffic is expensive to win. If only one in a hundred visitors converts, improving that to two in a hundred has the same effect as doubling your traffic, without the extra cost. This guide explains how to calculate and benchmark conversion honestly, a step-by-step CRO process, what to look for on key pages, how to run valid tests, which common mistakes to avoid and how to keep it ethical. For the services we offer, see our CRO services. For tracking foundations, see our GA4 guide.

What conversion rate is, and how to read it

Conversion rate is the share of visitors who complete a goal: conversions divided by visitors, times 100. If 2,000 people visit your service page and 50 submit the enquiry form, the conversion rate is 2.5 percent.

Goal type	Examples
Macro conversion	Purchase, booked appointment, qualified enquiry, trial sign-up
Micro conversion	Newsletter sign-up, add to basket, click to call, download

Important cautions:

• Benchmarks mislead. Published averages vary by industry, traffic source, device and price. Your own history is the best comparison.
• Segment the data. A site-wide rate hides big differences between mobile and desktop, new and returning visitors, and different traffic sources.
• Measure what matters. More form fills of low quality may not help. Track qualified leads and revenue, not just submissions.
• Count correctly. In Google Analytics 4, mark the events that matter as key events, and check they fire once per genuine action.

The CRO process in seven steps

1. Set up measurement. Confirm your key events and revenue are tracked accurately.
2. Find where people drop off. Use funnel and page reports to see where visitors leave.
3. Find out why. Combine data with qualitative evidence: recordings, surveys, interviews, support questions.
4. Form hypotheses. State what you will change, why you think it will help and what you expect to happen.
5. Prioritise. Rank ideas by likely impact, confidence and ease.
6. Test or implement. A/B test where traffic allows, or make careful changes and compare before and after, recognising the limits.
7. Learn and repeat. Record results, including failures, and feed them into the next round.

Step 1: Measure properly

Poor tracking makes CRO meaningless. Check that:

• Conversion events fire on the right action, once, with correct values.
• Forms and checkouts are tracked across the whole journey, including errors.
• Calls and chat are tracked if they matter.
• Internal traffic is excluded.
• Consent banners do not break measurement without you knowing.
• Sales or lead quality is recorded in your CRM, so you can judge outcomes beyond form fills.

Step 2: Find the leaks

Report	What it shows
Landing page performance by source and device	Where traffic arrives and which combinations convert poorly
Funnel or path exploration	Where users abandon a multi-step journey, such as basket to checkout
Form analytics	Which fields cause drop-off or errors
Site search terms	What visitors look for and cannot find
Page speed and errors	Slow or broken pages that cost conversions

Focus on pages with high traffic and high drop-off first. A small gain there is worth more than a big gain on a page few people see.

Step 3: Understand why, with qualitative evidence

Numbers tell you where. People tell you why.

• Session recordings and heatmaps. Watch how visitors scroll, click and hesitate. Look for repeated confusion, not single odd sessions. Respect privacy and mask sensitive fields.
• On-page surveys. Ask "What nearly stopped you from enquiring today?" or "What were you hoping to find?" at the right moment.
• Customer interviews. Ask recent buyers what they compared, what worried them and what convinced them.
• Sales and support teams. They hear objections daily.
• Reviews. Yours and competitors'.
• Usability tests. Ask five people to complete a task and watch where they struggle.

What to look for on key pages

Landing and service pages

• Message match. Does the headline continue the promise of the ad or search result?
• Clarity. Within five seconds, can a visitor tell what you offer, who it is for and what to do next?
• Proof. Reviews, case studies, credentials, guarantees, clearly shown and genuine.
• Price or range information. Hiding all cost information often creates doubt.
• One main call to action, visible without hunting, especially on mobile.
• Objection handling. Address the top doubts in the page itself, such as an FAQ.

Forms

• Ask only for what you need. Each extra field can reduce completions.
• Use clear labels, helpful error messages and mobile-friendly inputs.
• Explain what happens next and how soon.
• Offer alternatives, such as a phone number, for people who prefer to call.

Ecommerce product and checkout pages

• Clear images, sizes, delivery costs and returns information shown before checkout.
• Guest checkout and a minimum of steps.
• Trusted payment options and visible security cues.
• No surprise costs late in the process.
• Recovery emails for abandoned baskets, sent in line with consent rules.

Mobile experience

• Fast loading. See our technical SEO checklist for Core Web Vitals thresholds.
• Readable text, spacious tap targets and no intrusive pop-ups.
• Tap-to-call and simple forms.

Writing hypotheses and prioritising

A good hypothesis has three parts: the problem you observed, the change you will make and the effect you expect. Example: "Recordings show mobile visitors scrolling past the form without noticing it. If we move the form higher and add a button in the first screen, more will start it, increasing enquiries."

Prioritise using a simple score from 1 to 5 for each of three factors, then rank by total.

Factor	Question
Impact	If it works, how much will it matter?
Confidence	How strong is the evidence behind the idea?
Ease	How quickly and cheaply can we do it?

Running valid tests

A/B testing compares two versions at the same time with randomly split traffic. Done well, it separates real improvement from chance. Done badly, it produces false wins.

• Test one clear change at a time where possible, so you know what caused the result.
• Decide the success metric first. Do not choose after seeing the results.
• Run for full weekly cycles so weekday and weekend behaviour are both included.
• Do not stop early because the numbers look good on day three.
• Have enough traffic. Small sites often cannot reach reliable results for small changes. In that case, make bigger changes based on strong evidence and track outcomes carefully, accepting less certainty.
• Check for tracking errors before trusting a result.
• Segment results carefully. A win on mobile and a loss on desktop may need different decisions.
• Record everything, including inconclusive tests.

Be aware that testing tools that change page content with scripts can affect page speed and, if implemented badly, how search engines see the page. Google says pages must be indexable and consistent for crawlers, so avoid practices that show search engines something different from users.

A worked example

This example is invented. A Manchester dental practice gets 3,000 monthly visitors to its implant page and 30 enquiries, a 1 percent rate.

1. Measure. The team confirms form and click-to-call events are tracked, and records which enquiries become consultations.
2. Find leaks. Mobile converts at half the desktop rate. The form is below a long introduction.
3. Understand why. Recordings show mobile users scrolling and leaving. A survey reveals people want to know the cost before calling.
4. Hypotheses. A clear price range, a shorter introduction and a form near the top will increase enquiries.
5. Test. They test the new version against the old for four weeks.
6. Result. Enquiries rise, and consultation bookings follow, but the gain is modest and the team notes that some visitors still had questions about recovery time.
7. Next. They add a short recovery section and test again.

CRO and SEO working together

CRO and SEO support each other. SEO brings the right visitors, CRO turns them into customers. Be careful with changes to well-ranking pages. Keep the content and headings that earn traffic, and add conversion elements around them. Test on pages where you can afford to learn. See our SEO guide.

Keeping CRO ethical

• No fake urgency or scarcity. Countdown timers that reset, false stock warnings and invented "viewing now" counts mislead customers and can breach consumer protection rules.
• No hidden costs. Show the full price.
• No manipulative design that makes cancelling or declining hard.
• No fake social proof. Do not invent reviews or user counts.
• Be clear about consent and data use.

Short-term lifts from manipulation tend to harm trust, returns and reputation.

Common CRO mistakes

• Copying "best practice" lists without checking your own data.
• Testing tiny changes on low-traffic pages.
• Stopping tests early.
• Optimising micro conversions that do not lead to revenue.
• Ignoring lead quality.
• Skipping qualitative research.
• Running many overlapping tests that contaminate each other.
• Believing a single average conversion rate describes your site.
• Forgetting that faster, clearer pages often beat clever tricks.

A worked example: an ecommerce checkout

This is an invented example. A UK online shop selling craft supplies converts 1.8 percent of visitors. Mobile converts at 1.1 percent, and most abandoned baskets are on mobile.

Find the leak

Funnel reports show that 46 percent of mobile shoppers who reach the checkout page leave before entering their details. Desktop loses 28 percent at the same step.

Understand why

• Session recordings show people on mobile scrolling up and down the checkout page, looking for delivery costs, which are only shown on the final step.
• An exit survey asks "What stopped you completing your order?" Common answers: "delivery cost surprised me", "I had to create an account", "the page was slow".
• Page speed tests show the checkout takes over four seconds to become usable on a mid-range phone, partly because of third-party scripts.

Hypotheses

1. Showing estimated delivery costs earlier will reduce surprise and abandonment.
2. Offering guest checkout will reduce friction for first-time buyers.
3. Removing two non-essential third-party scripts from the checkout will improve load time.

Prioritise

Scoring impact, confidence and ease, the team ranks the speed fix first (easy, solid evidence), guest checkout second (easy, strong survey evidence) and delivery cost display third (more work, strong evidence).

Test and learn

The team removes the scripts and launches guest checkout together, since both are low risk and strongly supported by evidence, and compares the following four weeks with the previous four, noting that this is a before-and-after comparison, not a controlled test. Mobile checkout completion improves. It then runs an A/B test on delivery cost display for six weeks, with traffic split evenly. The test shows a smaller improvement than hoped, and the team notes that returning customers, who already knew the costs, diluted the effect. It records the result and moves to the next issue: product page photography.

A landing page teardown checklist

Area	Questions
First screen	Can a stranger tell what this is, who it is for and what to do next within five seconds?
Message match	Does the headline match the promise in the ad or search result that brought them?
Offer	Is the offer clear and honest, with price or a range where possible?
Proof	Are there genuine reviews, case studies, credentials or guarantees near the call to action?
Objections	Does the page answer the top three doubts? (Price, risk, time, trust, fit)
Call to action	Is there one main action, with clear wording, visible on mobile without hunting?
Form	Are there only the fields you need, with clear labels and helpful errors?
Speed and stability	Does it load quickly and stay stable on a mid-range phone?
Accessibility	Can a keyboard and screen reader user complete the action? Is contrast sufficient?
Next step	Do people know what happens after they submit?

How much traffic do you need to test?

The smaller the change you want to detect, the more visitors you need, and the lower your baseline conversion rate, the more you need. As a rough illustration, a page converting at 2 percent needs many thousands of visitors per version to reliably detect a modest relative improvement, and tens of thousands for a small one. Exact numbers depend on your confidence settings, and testing tools include calculators. If you do not have that traffic:

• Test bigger changes that you expect to have larger effects.
• Test on your highest-traffic pages.
• Combine qualitative research and sensible changes, then monitor results over a longer period.
• Test at an earlier step with more volume, such as add to basket, with caution about whether it reflects final sales.

CRO for lead generation businesses

For service and B2B firms, a form submission is only the start. Track what happens next.

• Define a qualified lead with sales, and record it in your CRM.
• Measure the lead-to-customer rate by landing page and source.
• Use a qualifying question on the form, such as budget range or timeframe, to reduce poor-fit enquiries, while balancing the effect on completions.
• Speed up the reply. Fast responses to enquiries often convert better. Track time to first reply.
• Offer more than one way to get in touch: form, phone, chat and booking calendar.
• Test the call to action wording, such as "Get a quote" versus "Book a 15-minute call".

Questions people ask about CRO

Question	Answer
What is the first thing to fix?	Tracking accuracy, then the highest-traffic, highest-drop-off page or step
Do button colours matter?	Rarely on their own. Clarity, contrast, wording and the offer matter more
Is longer or shorter copy better?	It depends on the decision. High-consideration purchases usually need more information, simple actions need less. Test with your audience
Should I add pop-ups?	Use sparingly. Intrusive pop-ups harm experience, and Google's guidance on page experience asks that main content be easy to find
How do I avoid harming SEO?	Keep content that earns traffic, avoid slowing pages, avoid cloaking and make sure testing tools do not hide content from search engines
What about live chat?	It can help answer doubts, if staffed or well designed, and clearly labelled when automated

A CRO reporting template

Field	Example
Test or change	Moved the quote form above the fold on the service page
Hypothesis	Mobile visitors were not noticing the form, so moving it up will raise submissions
Evidence behind it	Recordings and a heatmap showed visitors leaving before the form
Metric	Qualified enquiries per 1,000 visitors
Dates and traffic	Four weeks, 6,200 mobile visitors per version
Result	Improvement with a stated range of uncertainty, and any segment differences
Decision	Roll out, roll back or retest
Learning	What we now believe, and what to test next

A trust and friction audit for any key page

• Trust. Are there real people, reviews, guarantees, accreditations and clear contact details near the action?
• Clarity. Can a stranger explain, in one sentence, what the page offers after five seconds?
• Effort. How many steps, fields and clicks stand between interest and completion?
• Risk. Is it clear what happens if the customer changes their mind?
• Distraction. Are there links and offers that pull attention from the main action?
• Technical. Does it work on common phones and browsers, and with a keyboard and screen reader?

A story of a test that lost, and why it was still useful

This is an invented example about a Bristol online shop selling climbing gear. The team believed that adding a prominent countdown timer to its sale banner would raise conversion. They had read case studies about urgency, and a competitor used timers. They designed an A/B test: half of visitors saw the banner with a countdown, half saw the same banner without it, for three weeks across all sale pages.

At the end of the test, conversion was slightly lower with the timer, though not by a margin that proved a real difference. More tellingly, the customer service inbox showed an increase in questions such as "Does the sale really end then?" and "Will you extend it?" The shop's honest answer was that the sale would end on the stated date but could be repeated later in the year, which made the timer feel less credible. A few customers commented that they felt pressured. Some wrote that they would wait for the next sale.

The team could have treated the result as a failure and moved on. Instead, it wrote up what it learned. First, the idea of urgency was not wrong in itself, but the way it was presented clashed with the shop's brand, which was built on trust and honest advice. Second, the questions to customer service showed what visitors really wanted to know: is this a good price, and is it right for me? Third, the shop realised that it had no clear information on sale pages about why products were discounted, what was changing in the next range or how the sale pricing compared with usual prices.

The next test addressed those questions. The shop replaced the timer with a short, factual note near the top: "Last season's colours, reduced because new ranges arrive in March. Sizes are limited, and we will not restock these colours." It added clear before-and-after prices and a size availability indicator that reflected real stock. Conversion rose, returns were unchanged, and customer service questions fell. The shop noted that it could not be sure why, and that the test ran during a single season, so it planned a repeat.

The story illustrates three principles. Tests that lose are valuable when you look for the reason. Ethical, honest presentation is often also effective. And qualitative evidence, such as customer questions, can point to better hypotheses than best-practice lists. A CRO programme that records and learns from losing tests will, over time, outperform one that only celebrates wins.

Where we can help

We are a digital marketing agency in Manchester, UK and Mumbai, India. Our conversion rate optimization service starts with a tracking check and a research phase, and we report on enquiries and revenue, not just rates. Contact us for a review of your key pages.