Users are seldom as loyal as you think. Check your statistics!

What does it mean if a user visits your site often? Is it always a positive thing? Do frequent visits to a website represent a loyal user or customer? The latter is a common assumption, but consider the following statement from the web team of a cable-TV provider:

“We have loyal website users that keep coming back, so they must like our site and our company.”

A customer of the same company (which happens to be me …) sees it like this:

“They keep having service outages, and it takes so long to get through to customer support on the phone that I always check the website first to see if they already know about the problem.”

The slightly disgruntled customer who keeps coming back to the site is wrongly interpreted to be a happy customer who keeps coming back because she likes the website.

What is a reasonable frequency for user visits?

You are probably thinking, ”Surely, no website owner would make that wrong assumption?” Well, they do. All the time. At their peril.

Another example, this time from an electricity provider:

“We put new news items on our front page at least three times a week. That way our customers see that the website is up to date, and that something has changed since the last time they visited.”

The flip side of the coin, a customer of the same electricity provider:

“Usually, I just go to the website if there is something wrong, to find contact details, or to check my bill.”

The latter is typical for utility providers (e.g. electricity companies, gas companies, phone companies, etc.). If everything is running smoothly, there is little reason for existing customers to visit the website often. Customers get their bill monthly, or even quarterly, and for many customers visits as infrequent as montly, quarterly or even less frequent is quite probable.

This website team used a lot of resources on producing news items. A user visiting only two or three times a year will only see a few of these news items, and different users will see different news items. So the website is not a reliable way of communication news items to these users.

Revisits fuelled by user needs, not your website

Users will come to your site motivated by their own situation and needs. However, you can influence them in other channels – e.g. in ads, in newsletters, by providing RSS-feeds of relevant articles, referring to the website when people call you, referring to your logged in section on the website when you send them their bill, etc.

If the users’ needs and what they perceive that your website can provide coincide, they might visit your website again. So, in order to interpret what it means that your users visits you often or rarely, you have to look at other data and events besides what is happening on your site.

A local council experienced increased numbers of returning visitors in connection with the onset of influenza AH1N1, and the widespread vaccination of the population. The vaccinations where gradually carried out, region by region, and population group by population group.

The website had information and updates about the vaccination plans, as the information came in and as the vaccination doses became available in different regions. Many news sources referred to the local council website, and advised people to check the site regularly to see if it was their turn to be vaccinated.

Now, if the web team in the local council did not take into account the events in society as a whole, they might conclude that in 2009 the user loyalty of the website increased significantly. However, after the vaccination period was over, the revisits died down again.

This is an example of how events outside of the website team may give increased number of visitors, and revisits, but is not necessarily an indication of increased loyalty. So, you have to dig behind the numbers and look at which content the users came looking for, and why.

Take-aways

If you take something away from this article, let it be the following: Web statistics measures, including so-called ”loyalty”-measures are tools for detecting symptoms. They are invaluable. Like thermometers. However, you need to dig deeper and think harder in order to come up with a remedy, or even decide whether something is actually wrong.

PS:
Most web statistics tools will give you a measure of how many times your users have been to your website in a particular period.  It is useful to know how these measures are calculated, and the Web Analytics Association offers the following definition:

Visits per Visitor: The number of visits in a reporting period divided by the number of unique visitors for the same reporting period.

Visits per Visitor is termed Loyalty (under heading Visitor Loyalty)  in Google Analytics. See Web Analytics Association Web Analytics Definitions (PDF, 293 KB) for other definitions of web analytics measures.

There are issues relating to accuracy of web statistics based on the ”unique visitor” measure, e.g. because cookies may be deleted and make a repeat visitor look like a new unique visitor.  However, there are ways of compensating, and this measure is still useful to look at more closely, even with inaccuracies in the data.

For a dissussion of sources of inaccuracies, and relevant questions to ask in order to uncover such inaccuracies, see p. 12 in Web Analytics Association Web Analytics Definitions (PDF, 293 KB).

I’ve always found Avinash Kaushik’s blog Occam’s Razor a good source on all things web analytics. Also, have a look at Web Analytics Demystified, Eric Peterson is at the moment offering his book on Web Analytics Demystified as a free download.

What are your favourite blogs on web analytics?

Get the newsletter

Sign up to receive a reminder when new content is added to the calendar. (You may unsubscribe at any time.)