TfL website – Analytics insight one year on

With our new website having launched just over a year ago, the Analytics Team within TfL Online have been looking over the traffic stats for the new site. This post offers an insight into the usage of the site, the traffic peaks and troughs, and how we can use this information to help us.

What is our average daily traffic to

We get about 715,000 visits to the site per day. These visits are made by 575,000 unique visitors and over the course of a day about 3,100,000 pages are viewed. A visit is like a user session and it’s the key metric we use to measure customer traffic. Unlike the unique visitors metric, if you come to the site more than once a day each separate visit is counted.

Average Daily Visits
Our average daily visits stats show that Friday is the busiest day for traffic to our website.

How has this changed since the new site launched?

Our traffic stats are on the up! Between April and June 2014, the first 3 months after the launch of the new site, we were receiving 645,000 visits per day on average.*

What’s our busiest day of the week?

Traffic remains fairly stable throughout the week, but Friday is our busiest day in terms of both visits and unique visitors.  On Fridays, we see many visits from both commuters and leisure travellers, with the result being higher than average traffic between 4-6pm as they make their travel plans.

What’s been our busiest day since the website launched?

In terms of visits, April 29 2014 was our busiest day since launch. This was the first full day of a 3 day tube strike, when we received 2 million visits and served an incredible 9 million pages. There are no prizes for guessing our quietest day since launch – Christmas day 2014 when we had 331,000 visits.

How does this information help us?

Our figures are relatively predictable. For example, on Monday 27 April 2015 between 12-1 pm, we received 41,924 visits and in the same hour the previous week 41,538 (see graph below). From an operational point of view, these patterns give us a reliable basis from which to manage expected load. For the seemingly unpredictable, we can use historical data from strikes and weather events to project traffic.

It’s also very easy to spot when something is happening on the site or on our transport network. For example, last week on Tuesday 21 April between 8-9am we received 64,500 visits compared to 49,500 the week before, and this tells us that there were issues with the morning commute on that day.

Hourly visits to the site
Data for hourly visits to the site is relatively predictable, helping us manage expected load.

And the insight?

Of course, increasing traffic doesn’t always indicate engagement or satisfaction with the website. Repeat usage, attitudinal measures and successful task completion should be used to enhance traffic data, and these will help us to spot issues and opportunities.


*Unusual peaks excluded

Leave a Reply

Your email address will not be published. Required fields are marked *