Mobile Site Speed Benchmark

How Fast are the UK Top 5,000?

Introduction

In our recent study Mobile Readiness & Progressive Web Apps, we collated case studies that looked at the effect of slow site speed on mobile device users. All of these studies indicated that slow mobile experiences had a detrimental impact on visitors and visitor performance metrics. Following on from this research we thought it would be helpful to understand speed in context by extending the research and creating a mobile site speed benchmark so that you can see where your site fits against your peers across a range of speed metrics.

This UK Mobile Site Speed Benchmark compares over 5,000 (specifically 5,010) of the top UK sites from the public dataset at The HTTP Archive including by sector, so that you can compare your website to your category-peers. HTTP Archive uses WebPage Test, by Google’s Patrick Meenan, and in this case incorporates a Lighthouse runtime set-up.

lighthouse-runtime

How to Get the Data

Here’s the Google Sheet of all 5,010 sites and their classification. You can also request access to all of the data by category. Just navigate to the sector category you are interested in and see the “Get Data” call-to-action at the bottom of the page.

Read on to find out about the UK wide benchmark scores, (per speed metric) or jump straight to the sector that you’re interested in using the text links below. First, a quick note on the classification by sector:

 

Sector Classification

We wanted to expand on our PWA & Mobile Readiness study so kept to the same categorisations. In this case as there were over 5,000 sites to manually classify we tried to automate the process using Majestic for topic information, and assigned one of our category classifications as a roll-up category for the Majestic topic. When the URL was unknown (to the team here), and a conflict was found in the URL name and Majestic topic, the site was flagged for manual review for an editorial assignment to category. In this exercise we found that some categories were too large and contained websites for so many organisations-in-common, that we could divide the categories further to give greater granularity. For example the category “General Services” in our PWA & Mobile Readiness study has been split to more specific services e.g. Utility Services, Communications Services.

 

Distribution of Sites

The HTTP Archive uses Alexa to determine top sites, and as there are over 5,000 herein we are confident that this is a statistically robust data sample against which to measure your website performance.

Here’s how the sites are distributed by category:

And here’s that in absolute numbers…

number-of-urls

 

Average Speed Benchmark for the UK Top 5,000 Websites

average-speed-benchmark-all-metrics

 

Average TTFB

Time to First Byte (TTFB) is a well-established performance metric and is used to measure the time from the client request, to receipt of the first byte of information. It’s a good indication of a server response time, though is a means to an end in technical marketing performance terms, where the goal is to positively impact a visitor.

average-ttfb

News & Media sites had by far the fastest average TTFB at 1.07 seconds, and Utility Services, which include water and energy companies, the slowest at 1.55.

 

Average renderStart

Start Render (renderStart) is defined as the time necessary for a first element to render on the page, despite its size or its importance for the user; it can be text, background colour or any other element.

average-renderstart

 

Average onContentLoaded

The Document Content Loaded Time (onContentLoaded) is defined as the time that the browser takes to parse the document and execute deferred and parser-inserted scripts, including the network time from the user’s location to the server. Often is the starting point for JavaScript framework execution.

average-oncontentloaded

 

Average onLoad

The ‘OnLoad’ metric can be defined as the time until the browser reaches the load completion for the main documents in the browser; this time can be extended as pages grow and load a lot of content that is not visible to the user or off the screen (below the fold) even if the user-visible content has long-since rendered.

average-onload-time

 

Average visualComplete

The Visually Complete (visualComplete) metric indicates the time needed for the viewport area to be displayed under its final form and is very dependable on user’s resolution, since the placement of the fold line is associated with it.

average-visualcomplete

 

Average fullyLoaded

Fully Loaded Time (fullyLoaded) is the time measured from the start of the initial navigation until there are 2 seconds of no network activity – after complete loading of the page.

average-fullyload

 

Average Speed Index

The Speed Index metric measures how quickly the page contents are visually populated (where lower numbers are better). It is particularly useful for comparing experiences of pages against each other (before/after optimizing, my site vs competitor, etc) and should be used in combination with the other metrics (load time, start render, etc) to better understand a site’s performance. It is expressed in milliseconds and dependent on size of the view port. (Source: WebPage Test)

average-speed-index