Page Speed and Conversions: Measure the Link

Most website owners accept that faster pages are better, but few can say how much speed matters for their own conversions. Without that answer, performance work competes poorly against new features and campaigns, and teams either over-invest in shaving milliseconds or ignore real problems. Measuring the link between page speed and conversion on your own site turns a general belief into evidence you can act on. This guide explains which speed metrics to use, how to collect field data from real visitors, how to compare conversion rates across speed bands without fooling yourself and how to run experiments that show cause and effect.

Why general claims are not enough

Plenty of case studies report that faster sites convert better. They are useful encouragement, but they come from other businesses, audiences and pages. Your visitors, devices, offers and checkout flows are different. What matters for prioritising work is how speed affects your users on your key pages, and whether improving it would change results enough to justify the effort.

Choose the right page speed metrics

Speed is not one number. Useful metrics include:

• Largest Contentful Paint (LCP): when the main content appears
• Interaction to Next Paint (INP): how quickly the page responds to user input
• Cumulative Layout Shift (CLS): how much content moves unexpectedly
• Time to First Byte: how quickly the server starts responding

These three Core Web Vitals are described on web.dev, which also recommends measuring them at the 75th percentile of page loads, segmented across mobile and desktop (web.dev: Web Vitals). Our guide to Core Web Vitals explains each metric and its thresholds in more detail.

Use field data, not just lab tests

Lab tools test a page in a controlled environment with set device and network conditions. Field data comes from real visitors. web.dev explains that the two can differ, sometimes considerably (web.dev: lab and field data).

Lab tests are excellent for diagnosing problems and checking changes before release. To connect speed with conversions, you need field data, because conversions happen in real visits on real devices and networks. Collect speed metrics from real users, often called real user monitoring, and record them alongside session or visitor identifiers so they can be joined with conversion data.

Connect speed data to conversions

To analyse the relationship:

1. Record speed metrics for each page view or session
2. Record conversion events, such as form submissions, sign-ups or purchases
3. Join them by session or visitor
4. Group sessions into speed bands, such as LCP under 2.5 seconds, 2.5 to 4 seconds and over 4 seconds
5. Compare conversion rates across bands, by device type and page type

Focus on key pages: landing pages, product pages and checkout or form pages.

Worked example: page speed and conversion bands

The table below shows a hypothetical analysis of mobile sessions on a lead generation landing page over one month. Figures are illustrative.

LCP band (mobile)SessionsForm submissionsConversion rateUnder 2.5 seconds12,0004203.5%2.5 to 4 seconds8,0002162.7%Over 4 seconds5,000951.9%All mobile sessions25,0007312.9%

The pattern suggests slower sessions convert less often, but it does not prove speed is the cause. Slower sessions may come from older devices, weaker networks or regions with different buying behaviour. The next step is to control for these factors or run an experiment.

Beware of confounding factors

Correlation between speed and conversion can be misleading because other factors affect both:

• Device type and age
• Network quality and location
• Traffic source, such as paid social versus organic search
• Returning versus new visitors
• Page type and content

Compare speed bands within the same device type, traffic source and page to reduce these effects. Even then, treat results as indicative rather than proof.

Run experiments to show cause and effect

The most reliable way to measure the effect of speed is an experiment:

• Improve speed for a randomly selected group of visitors and compare conversions with a control group
• Or, more cautiously, test a small, deliberate delay on a limited share of traffic to estimate the cost of slowness

Experiments need enough traffic and time to detect differences. Decide your success metric and duration before starting, and avoid stopping as soon as results look favourable.

Look beyond the first page

Conversions often happen several pages into a visit. A fast landing page followed by a slow pricing page or checkout can still lose visitors. Analyse speed across the whole journey, not just the entry page, and look at where visitors drop out. Interaction delays matter too: a form that responds slowly when people type or press submit can cost conversions even if the page loaded quickly.

Estimate the value of improvements

Once you have an estimate of how conversion changes with speed, you can estimate the value of moving sessions into faster bands. Multiply the expected change in conversions by the value of a conversion, and compare with the cost of the work. Present the estimate with its assumptions and uncertainty.

Prioritise speed work by impact

Use the analysis to focus effort:

• Pages with high traffic and high conversion value
• Device types where speed is poorest
• Metrics with the largest gaps from good thresholds

Common improvements include optimising images, reducing JavaScript, improving server response times and caching. Our guide to Next.js performance covers techniques for sites built with that framework.

Keep measuring after changes

After each improvement, check:

• Whether field metrics actually improved, not just lab scores
• Whether conversion rates moved in the expected direction
• Whether other changes, such as campaigns or pricing, might explain differences

Keep speed metrics in regular reporting alongside conversion rates so regressions are noticed quickly.

Share results in business terms

Performance metrics mean little to most stakeholders. Present findings as conversion rates, leads or revenue by speed band, with clear notes on what is proven and what is estimated. A short summary showing which pages matter most, what improvement is realistic and what it might be worth makes it far easier to secure time for performance work.

Avoid common mistakes

• Relying only on lab scores
• Comparing averages that hide slow sessions
• Ignoring device and traffic source differences
• Claiming speed caused a change without testing
• Optimising pages that have little traffic or value

Frequently Asked Questions

Does page speed affect conversion rates?

It often does, but the size of the effect varies by site, audience and page. Measure it on your own site using field data and, ideally, experiments.

Which speed metrics should I use?

Core Web Vitals, including Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift, plus server response time, measured from real users.

Why do lab and field speed results differ?

Lab tests use controlled devices and networks, while field data reflects real visitors' devices, networks and behaviour, which vary widely.

How can I prove speed improvements increased conversions?

Run a controlled experiment comparing faster and standard experiences for randomly assigned visitors, with a predefined metric and duration.

Know what speed is worth to your site

We connect real user speed data with conversion data, run experiments and prioritise the performance work that pays. See our website conversion optimization service, or talk to us about page speed and conversions.

About the Author

Manav Sharma is the founder of Insyrge and a CRM solutions architect. He works across Zoho, Salesforce, HubSpot, Odoo and more, designing the CRM setups, integrations and automations that keep sales and operations teams working from the same data.