If your top sales representative told you he had made 50 sales calls last week, but it turns out he made 30, you’d have obvious cause for concern.  But did you know there’s a fair chance of the same thing happening with your electronic, 24/7 sales rep? Your website?  Because in the past few weeks I have taken a look at statistics gathered from a number of companies websites that simply do not add up, and the results are just plain scary.  Sometimes Lotto balls would seem to provide better numbers.  As it is a case of basic arithmetic and not rocket science, I want to share my concerns.  

And before you groan,  this isn’t a Red Sheriff sales pitch.  Sure, those of you familiar with Red Sheriff will know that accurate site measurement is our forté.  As a company we have spent the last five years developing one of the world’s most exacting website monitoring systems.  We record all cached pages.  Users. And exclude junk such as the traffic generated by search engine robots.  But the majority of companies still analyse their server logfiles and if they don’t use Red Sheriff I’d still rather see website managers using these in a better manner than nothing at all.

So one of the first things I do in meeting a new client is to check out what sort of website monitoring software they are currently using – in most cases the statistics are provided as part of an overall management package by the website’s host company.  Unfortunately, more often than not, the information provided is downright misleading.

What am I talking about? Well, with server logfile analysis the problems can be many, but two major issues usually glare at me from these reports, because they’re easily managed yet highly influential on the end result.  Firstly a large number of web pages are constructed from several frames.  Commonly one frame may be your tool bar across the top, the second the site directory down the side and the third the constantly changing body of information in the centre.  Many website statistics packages report each of these frames as a page unto itself, so when checking these statistics I often see individual page entries with descriptions such as “www.domain.co.nz/topbar.html” or “www.domain.co.nz/bottombar.html or even “www.domain.co.nz/leftnav.html”. 

So while your user has just clicked onto one page it is actually being reported as three separate pages and you immediately triple the apparent number of pages being accessed on your site.  Consequently site navigation frames frequently pop up in listings of the “Top 10” pages. My advice is to carefully read the page URLs and only count those that represent non-recurring information, remembering that other information such as “pages per session” remains compromised – but hey, you can’t have it all.

Secondly, another common report on log file packages is the “user organizations” page, listing where your site visitors come from.  This also needs careful reading.  Very often you will find substantial entries accredited to search engines.  These obviously should not be counted.  Search engines automatically, electronically access your site all the time.  It is not a human visit but simply the search engine doing its job.  Most of the time their identity is straightforward – no prizes for guessing which search engine uses the “Googlebot” programme.  Or, check the “browsers used” report for the “Mozilla” search engine browser equivalent. Similarly admin pages that are only viewed by the webmaster are also generated and often counted as your site being accessed – take these figures away and your are beginning to get closer to an accurate count. 

On the face of it, such an erosion of the traffic volume, which occurs when such “junk” traffic is excluded, can be a body blow to someone who has little else to justify the site’s existence.  But fortunately mere traffic volume has ceased to be the KPI it once was, and cutting out rubbish stats will definitely improve other measures – online conversion and sales ratios, for starters.  Which is why you need to start looking beyond simple visits and start focusing on what your online visitors are actually doing while they are visiting – better to have a hundred bona-fide customers than thousands of electronic tyre kickers. Online surveys can be a brilliant addition to website statistics for fuller, more informative insights.

At the end of the day, as a business owner or manager, you will be basing some of your e-strategy decisions on the statistical information you get from your website.  Even if you can only afford a basic package, if the information doesn’t add-up you can’t possibly make a good decision.  So take my advice and read those reports carefully.  You need to make sure the information you are using is as accurate as it can be and you don’t need to be a rocket scientist to do it. 

Jonathan Dodd