In this digital age web analytics, the measurement and reporting of meaningful metrics, is crucial to the performance of every 21st century business.
Equally crucial is that anyone whose job is to act on the resulting reports has a clear understanding of both the report and of how the metrics are defined. This is as much the case for web analytics as it is for any other domain. Any failure of understanding can lead to a serious failure in understanding customers and worst of all, in addressing any issues your customers are facing.
Let me give you a simple example.
Recently, as I was collecting some pre-paid tickets for a train journey, the machine display advised it was printing item 1 of 3, 2 of 3 and eventually 3 of 3. The machine dispensed a receipt and seat reservation, but no ticket. A brief visit to the ticket office to try and explain the situation to customer services compounded the problem.
I was told, ‘You must have left the ticket behind. Customers are always doing that (‘sigh!’). The staff gladly produced a report showing the printing record for the machine to demonstrate that all three items had been printed. Despite my protestations that the machine had ‘eaten’ my ticket I was told ‘…the printout ‘proved’ the ticket had been dispensed’. Note the subtle difference between, and interpretations of, the definitions of print and dispense. I tried to explain this without too much success!!
My experience got me thinking about clarity in describing what is actually being shown in a report. I’ve seen the same in customers and prospects in dealing with web analytics reporting.
My situation brought into focus the importance of measuring and reporting on the correct metrics for the task in hand and, perhaps crucially, developing an understanding by the report’s intended audience of the limitations in metrics and definition of the data being collected.
For example, we must be clear on the differences between a Visitor, a Visit, New Visitors and Repeat Visitors. Most web analytics software adopts the convention that a Visit will be deemed to have ended after 30 minutes of inactivity. In some cases this is a parameter that can be modified but for consistency and benchmarking across site properties it is usually left as the default.
Defining a Visitor in Web Analytics
If, for example, a Visitor to your web site enters the site for the very first time, and then comes back some time later that day or does not click on anything more and 30 minutes later makes their next click, they will be counted as both a New Visitor and a Repeat visitor on that day and that Visitor will have two visits.
So let’s say we have a loyalty report that tells us how many Visits our Visitors have had in a month, split by New Visitors, Repeat Visitors and Total Visitors. For Visitors who have only ever had one visit it’s easy, New Visitors+ Repeat Visitors = Total Visitors, because Repeat Visitors =0. But what about those Visitors with more than one visit to the site?
It’s a common question as the numbers don’t always appear to add up. But as we have seen earlier, it’s all a question of understanding the precise definition. A New Visitor in the time period being reported on has always made their first visit during that period.
However, a Repeat Visitor has visited the site before at some point possibly before or during the reporting period. So if the Visitor has had two visits during the reporting period they could be a New Visitor (for their first visit) and crucially a Repeat Visitor (for their subsequent visit). So now New Visitors + Repeat Visitors may not equal Total Visitors.
Real People Vs Visitors in Web Analytics
Is a Visitor the same as a Person? Again the recipient of the report needs to understand the definitions because Visitor does not mean Person. The most reliable identification method is an explicit login where the user identifies themselves by an identifier such as a username or email address. That way even if they arrive on the site from different devices, or browsers we can still identify them. A slightly less robust, but the most common, method is to use cookies which will identify repeat visits from the same machine and browser provided cookies haven’t been deleted.
So make sure your reports have clear and unambiguous definitions and usage guidance.
Oh, and in case you wondered about the tickets. After several requests at the ticket office, and after initially being told it was impossible, the frontline staff agreed to get an operative to open the machine to prove to me that the machine couldn’t have failed to produced a ticket and that it wasn’t hiding in the machine. And there, stuck in the machine, was my ticket, and about five others.
All’s well that ends well, as they say.
