Conversion Rate Optimisation: Where To Start

A practical guide from Higher Ground

Most businesses do not need more website traffic as urgently as they think.

They need to understand what happens to the people already arriving. Are visitors finding the right product? Do they understand the service? Can they complete the next step without becoming confused or losing confidence?

That is where conversion rate optimisation starts. Not with a brighter button, a pop-up or an A/B test, but with a clear commercial question:

What do we need more customers to do, and what is currently stopping them?

We have worked on conversion journeys for ecommerce stores, complex B2B websites and software platforms. The details vary, but the most effective CRO projects follow the same principle: gather evidence, identify the most valuable problem, improve the experience and measure what happens next.

This guide explains how to do that without turning CRO into an endless list of disconnected tactics.

What is conversion rate optimisation?

Conversion rate optimisation, usually shortened to CRO, is the process of helping more of the right people complete a valuable action on a website or digital product.

That action might be:

• Buying a product
  
• Submitting a qualified enquiry
  
• Booking a consultation or demonstration
  
• Requesting a quote
  
• Starting a software trial
  
• Completing software onboarding
  
• Renewing or upgrading a subscription
  

The important phrase is valuable action. More form submissions are not helpful if they are poor-quality leads. More trial registrations mean little if customers never activate. A higher ecommerce conversion rate can still be commercially weak if discounting reduces margin or increases returns.

Good CRO connects user behaviour with business value. It improves the experience for customers while producing better commercial results.

Where should you start with CRO?

Start by defining the outcome that matters most, then work backwards through the customer journey.

Do not begin with a list of website changes. Begin with the problem you are trying to solve. For example:

• We receive plenty of enquiries, but too few are commercially relevant.
  
• Customers view products but do not add them to their basket.
  
• People begin the booking journey but leave before payment.
  
• Software users register but do not complete the first important task.
  
• Paid search generates traffic but not enough revenue to justify the spend.
  
  

Each problem requires different evidence and a different solution. This is why copying a competitor's website or following a generic CRO checklist rarely produces a reliable strategy.

1. Define the conversion and its commercial value

Before opening Google Analytics, agree what success actually means.

Most websites have one primary conversion and several smaller actions that support it. An ecommerce store may use completed purchases as its primary conversion, while product views, add-to-basket actions and checkout starts help reveal where customers are progressing or leaving.

A B2B website may generate brochure downloads, email clicks and contact forms. But the real commercial outcome could be a sales-qualified enquiry or signed contract several weeks later.

Ask these questions first:

• Which customer action creates genuine value for the business?
  
• What is an average conversion worth?
  
• Which types of lead or customer are most valuable?
  
• Which smaller actions indicate progress towards the main outcome?
  
• Can website activity be connected to sales, revenue or retention?
  

This gives the CRO work a commercial baseline. Without it, teams often optimise whatever is easiest to count rather than what is most valuable.

2. Check that your tracking can be trusted

Analytics is useful only when it records the right actions accurately.

Before drawing conclusions, check that your analytics setup is collecting meaningful events and that important actions are marked correctly. Google now describes particularly important website actions as key events in GA4.

For an ecommerce store, that may include product views, add-to-basket actions, checkout stages and purchases. For B2B lead generation, it may include form starts, successful submissions, telephone clicks, booked meetings and qualified opportunities recorded in the CRM. Google publishes separate recommended events for online sales and lead-generation journeys.

Check for common measurement problems:

• Forms that record a conversion when they are opened rather than submitted
  
• Duplicate conversions caused by thank-you pages being refreshed
  
• Internal staff or spam traffic distorting reports
  
• Purchases recorded without the correct revenue
  
• Consent settings creating unexplained gaps in the data
  
• Website leads that cannot be connected to their eventual quality or value
  

You do not need a perfect analytics system before making any improvement. You do need to understand its limitations. Treat uncertain data as a clue, not an unquestionable fact.

3. Find the most valuable point of friction

Once measurement is reliable enough, look for the point where customer progress breaks down.

Do not simply start with the page that has the highest exit rate. People naturally leave confirmation pages, contact pages and articles after completing what they came to do. Context matters.

Instead, compare behaviour across the complete journey:

• Which landing pages attract commercially relevant visitors?
  
• Can they quickly understand the proposition?
  
• Do customers reach the pages needed to make a decision?
  
• Where do they hesitate, go backwards or repeat an action?
  
• Which devices, traffic sources or customer groups struggle most?
  
• Where is the greatest gap between customer intent and completion?
  

A small problem near a high-value conversion can matter more than a dramatic issue on a low-value page. Prioritise the combination of commercial impact, customer difficulty and confidence in the evidence.

4. Use quantitative and qualitative evidence together

Analytics can show you where something is happening. It rarely tells you why.

That is why CRO works best when quantitative evidence is combined with direct observation and customer feedback.

Quantitative evidence includes:

• GA4 journeys and key events
  
• Ecommerce and revenue data
  
• Form completion and lead-quality data
  
• Search performance and paid campaign data
  
• Device, browser and page-speed reports
  

Qualitative evidence includes:

• Session recordings and heatmaps
  
• On-site surveys
  
• Customer and sales-team interviews
  
• Usability testing
  
• Support queries, reviews and lost-sale feedback
  
  

A session recording may reveal somebody repeatedly opening delivery information. A customer interview may explain that the delivery date affects whether the order can be placed at all. Analytics may then show how many other customers leave at the same stage.

Even a small qualitative study can expose serious usability problems. Nielsen Norman Group recommends around five participants for a typical qualitative usability study involving one user group.

5. Check whether the proposition is clear

Many apparent conversion problems begin before the form, basket or checkout.

If visitors cannot quickly understand what you offer, who it is for and why they should choose you, changing the call-to-action button will not solve the real problem.

A strong page should answer:

• What is being offered?
  
• Who is it designed for?
  
• What problem does it solve?
  
• Why should the customer trust this business?
  
• What does it cost, or what determines the price?
  
• What should the customer do next?
  

The answers do not all need to sit in the first screen. They do need to appear at the right point in the decision.

Clear copy is not about removing every detail. A high-value B2B service or considered ecommerce purchase may need substantial information. The goal is to organise it around customer questions rather than internal departments or marketing slogans.

Our guide to copywriting for CRO explains how headlines, evidence and calls to action can support that decision.

6. Map the complete customer journey

Individual pages do not convert in isolation.

A customer might first see a paid advert, enter through a service page, read a case study, return through Google several days later and then submit an enquiry. An ecommerce customer may compare several products, check delivery terms, leave, see a remarketing advert and complete the purchase on another device.

Map the main stages from the customer's perspective:

1. Arrival: Does the page continue the promise made by the advert, search result or referral?
   
2. Understanding: Can the visitor recognise that the offer is relevant?
   
3. Evaluation: Can they compare options and find the information needed to decide?
   
4. Confidence: Is there credible evidence, pricing information and reassurance?
   
5. Action: Is the next step clear, proportionate and easy to complete?
   
6. Follow-up: Does the confirmation, sales response, delivery or onboarding fulfil the promise?
   

This often reveals gaps between marketing, website and sales activity. A well-designed form cannot compensate for a slow or unhelpful response after it is submitted.

7. Protect the brand while removing uncertainty

CRO is sometimes associated with aggressive pop-ups, countdown timers and pressure-led copy. Those methods may create a short-term response, but they can also make a premium or trusted brand feel desperate.

Branding creates recognition, preference and trust. UX makes the experience clear and usable. CRO measures whether the complete journey is helping more people act.

The strongest improvements remove uncertainty without removing personality. That might mean clearer product photography, better comparison information, more useful case studies, a simpler booking journey or a more honest explanation of what happens after an enquiry.

For luxury jewellery brand Maveroc, visual impact needed to work alongside product discovery, trust and a purchase journey appropriate for a considered luxury purchase. Brand and conversion were parts of the same customer decision.

8. Turn evidence into a testable hypothesis

A list of observations is not yet a CRO plan.

Each proposed improvement should connect evidence to an expected result. A useful hypothesis follows a simple structure:

Because we observed [evidence], we believe changing [part of the experience] for [customer group] will improve [commercial or behavioural measure]. We will know by measuring [result].

For example:

Because mobile customers repeatedly open delivery information and leave before adding a product, we believe showing the delivery date beside the purchase button will increase mobile add-to-basket rates. We will measure add-to-basket rate and completed orders without increasing cancellations.

This is far more useful than “make the delivery information clearer”. It defines the customer, the problem, the proposed change and the measure of success.

9. Choose the right way to test the improvement

A/B testing is useful, but it is not the starting point for every business.

A statistically reliable split test needs enough relevant traffic and conversions. A low-traffic B2B website may wait months for a meaningful result. In that situation, usability testing, customer interviews, prototype testing and before-and-after commercial measures can produce more useful evidence.

Choose the method to suit the question:

• A/B testing: Compare two versions when traffic and conversion volumes are sufficient.
    
• Prototype testing: Test a new journey before paying for development.
    
• Usability testing: Observe whether representative customers can complete important tasks.
    
• Five-second testing: Check whether the proposition and first impression are understood.
    
• Before-and-after analysis: Monitor behavioural and commercial changes when a split test is impractical.
  

The purpose of testing is not to prove the team was right. It is to reduce uncertainty and learn what should happen next.

10. Treat CRO as continuous improvement

A one-off audit can uncover important problems, but customer behaviour, marketing campaigns, products and competitors continue to change.

The most effective approach is a continuous loop:

• Gather insight‍
• Prioritise the next problem
• Design an improvement
• Test or implement it
• Measure the commercial result
• Use what you learn to choose the next priority
  

This does not mean constantly redesigning the website. It means making focused improvements instead of waiting several years for another large rebuild.

For ecommerce brand APOC, CRO insight and design refinement helped triple the conversion rate while preserving the store's identity. Continued Webflow, UX and CRO improvements for Fettle contributed to a fourfold increase in conversions as the business grew.

For JCB, insight-led changes to its website templates produced a 66% increase in new machine enquiries and a 300% increase in large excavator leads during the first three months. The templates were subsequently adopted across more than 12 countries. The point was not simply to generate more form submissions, but to improve high-value demand.

How CRO changes for B2B, ecommerce and SaaS

The method is consistent, but the definition of success changes.

B2B lead generation

Focus on qualified enquiries rather than form volume alone. Connect website sources and pages to CRM outcomes wherever possible. Clear propositions, sector evidence, case studies and an appropriate next step usually matter more than reducing every form to two fields.

Ecommerce

Measure product discovery, add-to-basket behaviour, checkout completion, average order value, margin, returns and repeat purchases. Product information, photography, delivery, stock, sizing and returns can influence conversion as much as page layout.

SaaS and software

Registration is not the end of the journey. Measure whether customers complete onboarding, reach their first useful outcome, adopt important features and remain active. CRO and product UX need to work together.

A practical CRO starting checklist

• Choose one valuable conversion to improve.
• Confirm that it is being measured accurately.
  
• Give it a commercial value or lead-quality measure.
  
• Map the steps customers take before and after it.
  
• Use analytics to find the largest point of friction.
  
• Use recordings, interviews or tests to understand why it happens.
  
• Write one evidence-based hypothesis.
  
• Choose a suitable test or implementation method.
  
• Measure the result and decide what to improve next.
  

If you do only one thing, do not begin by changing the button colour. Choose one commercially valuable customer journey and watch real people try to complete it. The points where they hesitate, become confused or lose confidence will give you a much better place to start.

Start with a focused CRO review

Higher Ground combines UX research, analytics, design and conversion rate optimisation to improve ecommerce stores, lead-generation websites and software platforms.

We start by identifying the most valuable customer journey, gathering evidence and prioritising the changes most likely to improve commercial performance. That may lead to clearer content, a redesigned journey, user testing, an A/B experiment or a programme of continuous improvement.

Read more about our CRO audits, explore why CRO should be part of your digital marketing strategy, or talk to Higher Ground about your website.