After fighting for years against page views, and for the use of time instead, let’s criticize now the concept of visits.

What's a visit? Well, that is the problem.

The visit on the Internet has no precise definition. Or more precisely, it has lots of definitions.

In the panels (user-centric tools), NetRatings and comScore use a period of inactivity of more than 30 minutes.
OK, but it's not so simple:

• Is the user looking at the computer screen?
• What if he is reading emails?
• And if the page is automatically refreshed?
• And if he watches a video for an hour?

Well, in the best case (panelists have installed a software that can measure everything), there are choices to make, and everybody is not doing the same choices.

In "site-centric" tools, it is even worse!
These tools only see the requests that were made on the measured site. So if the visitor leaves to another site and returns after an hour, there is no way of knowing whether the session was interrupted or not … It does not therefore correspond to the same visit measured by the panels (that track all the visited sites).

In addition, with our broadband connections, we use the Internet differently. Instead of connecting, surfing, disconnecting, we now make very short sessions (a few pages). Thirty minutes of inactivity is no more significant.

So today, the concept of visit means nothing.

But it is an interesting concept, though. But what is interesting, it is not the word "visit", but the idea of frequency.

The consumers panels, those measuring purchases, use three independent indicators to explain the total sales of a product:

• The number of buyers (1)
• The frequency of purchases per buyer (2)
• The average amount per purchase (3)

Multiply 1 by 2 by 3 and you get the total sales! And you have three ways to increase them depending on which indicator you act on.

On the Internet, you have the same concepts:

• Visitors (preferably unique)
• The frequency of visits (I did not say "the visits")
• The time spent per unit of frequency

And you can play on each indicator to increase the total time spent on the website:

• Recruit new visitors
• Update the website often so that they come more often
• Provide interesting content to be keep them longer.

The visit frequency is therefore also interesting on the Internet. So what definition shall we use?

At Alenty, we have chosen a simple concept: the day.

This notion is not impacted by the method of measurement. The definition may, at most, be refined by starting the days when there is less online activity (as it is done for television) at 5 am for instance.

Our tools provide thus:

• The number of visitors (as unique as possible)
• The number of days per visitor
• Time spent per day

Some tools acknoledge the difficulty of measuring unique visitors. So they use visitors instead. And what is a visitor (compared to a unique visitor)? A visitor is unique visitor on one given day. So, at worst, if you measure wrong VU, the number of days is automatically adapted and time per day remains right. And the total time spent on the site (UV x days x time per day) remains right too!

And we come back to the only truly stable and comparable information: the time spent!