Skip to article content

Chat, videos and other tools: are they slowing your website down?

A chat widget, map or video can add work to your website. Identify what loads, investigate its effect and assess changes without losing a useful function.

A useful tool may still need a performance review

You add a chat widget for enquiries or a video explaining a product. Later, the website seems slower to load or respond. The tool may contribute to that delay, but check its effect before removing it or treating it as the cause.

Some features load files and services from other businesses: video players, maps, bookings, reviews, analytics or advertising. These are external tools, also called third-party resources. Their usefulness and the work they add need considering together.

The web.dev guide to third-party JavaScript explains how these resources can affect loading and browser work. It does not establish that every integration is problematic or that removing all of them suits your business.

List the tools and where they appear

Follow customer journeys through the homepage, services, product details, contact page and checkout where relevant. Record which tools appear and what they do. Your technical team can add less visible resources, such as measurement tags or services loaded by another tool.

The number of providers does not establish their impact. One integration may request several files; several small tools may behave differently. A closed chat icon also does not establish that nothing has loaded.

Record information that helps decide what to investigate
ToolWhat to recordUseful question
Chat or customer supportPages where it appears, operating hours and owner.Does it need to load the same way across the website?
Videos, maps or reviewsLocation and when visitors need them.Could they be prepared when someone is about to use them?
Bookings or paymentsDependent journey and expected confirmations.How will an improvement be tested without breaking the transaction?
Analytics or advertisingPurpose, people using its data and consent management.Are tags duplicated or campaigns no longer running?

Add the provider, who can configure it and evidence of the delay. A clearly identified tool is more useful than technical names nobody can connect with a function.

Describe whether loading or a particular action is slow

A page can take time to show content, shift when an element appears or respond slowly when a button is pressed. These are different problems. The tool may contribute to one, several or none of them.

Record the page, steps, device and whether the delay occurs on entry, while scrolling or when opening a feature. If a cookie notice appears, note your choice: conditions may change which resources load. Compare equivalent choices while respecting people’s preferences.

If the page is already visible but a menu or filter responds late, our interaction and INP guide helps refine the observation. If its location remains unclear, start by identifying affected pages and tasks.

Arrange a controlled comparison with its context preserved

Your team can investigate a tool by comparing the same page and journey with and without the resource in a test environment or a narrowly scoped technical test. Record exactly what changed and what remained constant.

Repeat variable observations and retain reports. Changing hosting, templates and chat loading together makes attribution difficult. The linked web.dev guide describes technical resource comparisons; choosing the test and preserving the function also needs knowledge of your website.

Blocking a service during a test is not the same as preparing a solution for release. It may remove a booking function, prevent a purchase or change measurement. Record that loss of function rather than presenting a page that no longer does its job as a valid improvement.

Consider when and where it is needed before removing it

If tests connect a tool with the delay, options may include loading it only where needed, preparing it when someone uses it, correcting duplicate installation or considering an alternative. Not every tool supports the same options.

Discuss these choices with the people using the function. A chat supporting sales has a purpose; a map may help customers find the premises. The decision needs performance evidence and a way to retain that usefulness.

Review provider instructions, consent preferences and accessibility too. An alternative video presentation should make its purpose clear and allow keyboard use. Payments and bookings need their own tests and conditions rather than a blanket “load it later” rule.

An example: a video near the bottom of a page

Fictional situation without measurements or results attributed to a real website.

A product page includes a video after its description and specifications. The team suspects that loading the player adds work on entry. It prepares a comparison retaining the content while replacing the initial player load with a presentation that opens it when requested.

Checks would cover loading and interaction, but also whether visitors recognise the video, can play it correctly and can operate it with a keyboard. If the comparison shows no meaningful benefit, the change should not be presented as a demonstrated optimisation.

The example proposes a test. It does not establish that every video should use this arrangement or that a particular replacement will increase sales.

Keep a useful function and define an improvement you can check

At the end of the investigation, record the tool, purpose, test conditions, results and decision. Include what was preserved and what needs checking after release.

If no meaningful contribution is established, retain that conclusion and investigate another explanation, such as hosting capacity or configuration. An investigation can produce a useful answer without removing a valuable tool.

We can help review website performance and integrations within an agreed scope. The aim is to understand what deserves improvement and why before commissioning changes affecting customer support, measurement or sales.