Website hits as a performance indicator or performance measure

Hi.  Welcome to the forum.  

2GC has operated a well known and popular web site for over a decade now,  and during that time we've naturally been interested in finding out about visits to the site.  We use this information in a variety of ways, but mostly we use it to help us understand more about three areas of interest.

• First, we are interested in the raw level of activity on the site - we use this mostly to help us ensure the site server is not being overloaded.
• Second, we are interested in how activity on the site changes in response to changes we make in the site content and design - to help us work out if the changes we make are making the site more or less useful.
• Third, we are interested in anomalies - more than once we've found bugs in site content and programming by looking to understand oddities in the site's activity statistics.

Over the decade we've found that understanding these three areas requires us to look at a variety of performance measures, and to supplement this raw data with more direct / subjective feedback from people who actually use the site.  In doing this we are simply echoing (effectively we hope) what has become widespread best-practice in monitoring and learning about web site performance.  For us at least, 'website hits' are themselves a relatively unimportant number - we look more closely at what the field calls "visitors" (by collecting all the 'hits' from a particular web location in a given time window into a single 'visit'), visit duration (both in terms of how long the person stays on the site, and how many pages they look at and how long for), and visitor point of origin (we are an international firm, and so this is helpful to know).  We are also interested in which pages people look at most, and what 'path' visitors follow through the site to find whatever they are interested in.

What are the pros / cons?  The pros are pretty big really.  We've used our monitoring of the site traffic to improve the site (at least in so far as we have increased the number of visitors, the time they stay on the site, the speed they get to the page they want to get to, and their diversity in terms of originating country).  We've also been able to make informed decisions about where to develop the site etc.  The cons are smaller, but significant.  Over the decade we've used about four different performance tracking methods, each with their own complications and quirks - all have worked, but the need to learn about and implement each of them has triggered some tangible investment costs.  The systems we use now are better, but also more complicated than ever before.  Working out how to use the tools is hard, and for big commercial web sites has become a job specialisation in its own right.

In the light of all this, I would also, of course, encourage you to visit the 2GC web site that has been the focus of our attention.  The 2GC site has a diverse collection of Performance Management related content to review, and doubtless is of huge merit if you are interested in performance management as a more general concept, but sadly it lacks anything specific on website performance monitoring.  However, fortunately there are many other useful places to find out more about how to track and monitor website performance.  One resource I have found helpful is some material provided by IBM within its "Developerworks" system.  In particular a series of articles called Measuring Web traffic that you can find here.

Hope this helps.