How to use web analytics to optimise your website’s frontend? Part two: taking action and measuring results

Gathering relevant data is the first step towards improving your website. The next is to turn those insights into changes that make the site easier to use and support your business goals. 

In the first part of this article, we explored the tools and metrics that help teams analyse the frontend. Now we move into action: from reducing loading times and improving navigation and CTAs to checking whether those changes deliver the intended results. 

Data-informed optimisation 

Where should you start when the list of potential improvements seems endless? Analytics data and test results help identify which parts of the website need improvement. They provide a basis for planning work on performance, usability and conversion, then assessing what has changed for the better. 

Reducing page loading times 

Before visitors can explore your offering, they need to be able to see it. If your website keeps them waiting, it tests their patience from the outset. Fast loading makes the site easier to use and can support both engagement and conversion. 

Tools such as PageSpeed Insights, Lighthouse, WebPageTest and Chrome DevTools help identify what is slowing the website down. Depending on the findings, potential improvements include: 

  • Lazy loading – deferring the loading of images and embedded content outside the initial viewport. 
  • Minifying HTML, CSS and JavaScript – removing unnecessary characters and comments to reduce file sizes. 
  • Compressing resources – using methods such as GZIP or Brotli. 
  • Optimising file loading – removing unnecessary resources and loading the remaining ones when needed. 
  • Caching – configuring how resources are stored and reused. 

Each technique addresses a slightly different problem, so the choice should follow the diagnosis. Applied incorrectly, a technique can have the opposite effect. For example, lazy loading should not delay the download of the main image visible as soon as a visitor arrives on the page. 

Improving navigation and the purchasing process 

Path analysis, heatmaps and session recordings help identify where users encounter difficulties. These insights can inform changes such as making interface elements more intuitive, simplifying a form or purchasing process, and reducing elements that confuse users or interrupt their progress. 

Before redesigning a basket or form, however, it is worth establishing why people abandon it. The interface may be the obstacle, but the terms of the offer could be just as relevant. Data points to the area that needs attention, but the underlying cause still needs to be investigated. 

Optimising CTAs 

A call-to-action button may be a small element, but it can have a significant influence. Its visibility, wording and placement affect whether users take the next step. Optimisation may involve: 

  • checking whether users reach the button, 
  • improving contrast and readability,
  • testing different wording,
  • aligning CTA placement with the user journey,
  • reducing competing elements around it. 
Important!
Not every CTA needs to appear in the first screen visitors see when they open a page. What matters is that it appears in a logical position, at a point where users have the information they need to make a decision. 

Click analysis and session recordings help identify overlooked CTAs and develop hypotheses about why they are being missed. The effect of a change should also be assessed further along the journey, for example through completed form submissions rather than clicks alone. 

Personalising content 

Someone returning to your offering and someone visiting the website for the first time may need different information. Personalisation helps make the message more relevant to the context of the visit, taking into account factors such as: 

  • device type,
  • traffic source,
  • location or language,
  • behaviour during previous sessions. 

Data from GA4 and marketing automation platforms can help teams select personalisation scenarios, such as tailored messages, product recommendations or promotions. Displaying that content, however, requires suitable website functionality or an integrated tool. 

The aim is to deliver a more relevant message that makes it easier for visitors to take action. Whether personalisation achieves that aim needs to be measured, just like any other website change. 

Adapting the website to different devices 

The same website can offer very different experiences on a large monitor and a phone. Data on devices, screen resolutions and mobile journeys helps establish whether the layout and functionality suit the way people use the site. Potential improvements include: 

  • adjusting component layouts to suit the screen width, 
  • reducing unnecessary elements while retaining the information needed to complete a task,
  • making buttons, menus and forms easier to use on touchscreens. 

Getting the appearance right is only part of the task. A form may fit on a phone screen and still be awkward to complete. Testing should therefore cover the entire user journey, from reading the content to completing the most important action. 

A/B testing – how to check whether changes work 

A new headline may sound more persuasive, but will visitors agree? A/B testing allows teams to compare versions of a page or an individual element and assess their impact on user behaviour. This reduces the risk of implementing changes based on opinion alone. 

How does an A/B test work? 

An A/B test compares the current version, known as control variant A, with a modified version, variant B. Users are randomly assigned to the variants, and selected metrics are then compared, such as: 

  • conversion rate, 
  • CTA clicks,
  • progression to subsequent stages of a process,
  • engagement. 

The primary metric should reflect the purpose of the change. If the aim is to generate more enquiries, an increase in clicks on the contact button is not enough to confirm success. 

What should you test? 

CTA buttons 

A button’s wording, appearance and placement are natural starting points for testing. Their impact should be assessed through both clicks and progress towards the final goal. 

Headlines and introductions 

The opening lines help visitors decide whether to keep reading. Tests can reveal which way of presenting the offering is more effective at encouraging people to explore the content and take action. 

Section layout 

The order in which users encounter information also matters. Tests can compare section layouts, content length and the placement of visual elements. 

Forms 

The number of fields, their labels and the submit button deserve attention when users start a form but fail to complete it. Testing helps establish how changes affect the completion rate. 

Trust signals 

Reviews, client logos and certifications can help visitors feel confident about their decision. Tests can compare versions with and without these elements, as well as different placements. 

How should you analyse the results? 

Success criteria should be agreed before the test begins. A clear objective and a defined assessment method make the results easier to interpret. Once the experiment has ended, check: 

  • whether the sample size was sufficient,
  • whether the test ran for long enough,
  • whether the observed difference is statistically reliable, 
  • whether the result matters to the business. 

Experimentation tools, GA4 and BI dashboards can all support the analysis. Their roles should, however, be distinguished: a tool that allocates traffic and runs the test serves a different purpose from one that reports on user behaviour. 

Important!
What if the test produces no clear winner? That result still provides information. An inconclusive outcome does not automatically mean that the two versions perform identically: there may simply be too little data to detect a difference. 

Good practice for A/B testing 

  • Start with a clearly defined change. Testing one element makes it easier to attribute the result to a specific modification. 
  • Plan the sample size and duration. There is no single test length that suits every website; traffic volumes and conversion numbers matter. 
  • Account for unusual periods. Holidays and promotional campaigns can influence user behaviour and how the results should be interpreted. 
  • Record your findings. This makes it easier to plan future improvements and avoid repeating the same tests. 

A well-planned test brings structure to website development. Its greatest value lies in testing a hypothesis, including when the outcome differs from the team’s expectations. 

Recommendations for website owners and marketers 

Effective frontend optimisation brings together people responsible for business, marketing, design and technology. How can those perspectives inform the website’s day-to-day development? The following principles offer a useful starting point. 

1. Connect performance with business goals 

Website performance shapes the conditions in which users decide whether to make a purchase or get in touch. It should therefore be discussed alongside conversion, customer acquisition costs and campaign results. Regularly check whether technical issues are making it harder for the website to achieve its main goals. 

2. Use data to inform decisions 

Page layout, form length and CTA wording should reflect more than the team’s preferences. Analyse GA4 reports, heatmaps and session recordings, then use those findings to develop hypotheses and assess the effectiveness of potential solutions. 

3. Work closely with the technical team 

Website owners do not need to understand every implementation detail to play an informed role in the site’s development. What matters is knowing which problem a change is intended to solve and how success will be recognised. Agreeing priorities together helps connect technical work with users’ needs and business goals. 

4. Make optimisation a regular practice 

One successful release does not bring optimisation to an end. Make time for regular report reviews, further tests where traffic volumes allow, and documentation of changes, results and findings. This routine helps teams respond to emerging problems and check whether earlier improvements are still delivering the expected benefits. 

5. Review the mobile experience 

Analyse the share of mobile traffic and performance on phones. Test loading times, content readability and the usability of key features on smaller screens. Pay particular attention to the forms, menus and buttons that users rely on to complete the website’s main goals. 

6. Choose suitable tools and build the team’s knowledge 

An advanced BI system is not always necessary at the outset. GA4, a behavioural analytics tool, PageSpeed Insights and Lighthouse can provide a wealth of useful information. What matters most is the team’s ability to interpret the results and use them to plan its work. 

Conclusion: from analysis to a better-performing website 

Frontend optimisation begins with a close look at how people use a website. Technical diagnosis, implementation and evaluation follow. Every stage serves the same purpose: making it easier for visitors to take actions that also matter to the business. 

The greatest gains come when marketing and technical teams work towards solving a clearly defined problem and agree from the outset how the outcome will be assessed. This gives successive improvements a coherent direction and helps guide the website’s development. 

At Infinity Group, we bring together web analytics, UX and UI design, and frontend development. We help turn observations of user behaviour into changes that make websites easier to use. Want to understand what is holding your visitors back and where improvements would make a difference? Complete the form below, and let’s discuss how data can help improve your website.

    Contact us

    *Required

    Clause:

    The administrator of your personal data is Infinity Group Sp. z o.o., with its registered office in Białystok.
    The data provided in the form will be processed for the purpose of responding to your inquiry (Article 6(1)(f) of the GDPR – the administrator’s legitimate interest consisting in conducting correspondence). Providing your data is voluntary, but necessary in order to receive a response.
    You have, among others, the right to object to the processing of your data and the right to lodge a complaint with the President of the Personal Data Protection Office (Poland). Detailed information, including information on data recipients, the data retention period, and possible transfers of data outside the EEA, can be found under the link “Information on the processing of your personal data”.