Heal Your Church WebSite


Teaching, rebuking, correcting & training in righteous web design.

5 things we can learn from the SiteMeter debacle

Filed under “If it ain’t broke don’t Fix it,”SiteMeter, a webpage tracking service popular with pundit blogs, was compelled to roll-back a deployment and issue a public apology after they messed-up badly when they deployed new features that inflamed the blogosphere more than the myth that Palin and Obama are the pawns of an upcoming alien abduction to convert both Christians and Muslims to tenents of XENU. The end result were a number of prominent A-list political bloggers voicing dissatisfaction with the new SiteMeter reminiscent of the New Coke disaster of 1985.

A situation well described by  The Reference Frame’s LuboÅ¡ Motl,  who provides a detailed enumeration of the issues afflicting the new design, including:

  1. significantly slower due to Flash infusion
  2. fonts are too small, and not scalable
  3. does not associate the domain with IP
  4. requires left-to-right scrolling
  5. harder to quickly see the number of visitors to one’s site

What catches my eye are points #1 and #5 as they indicate to me the possibility that the good intentions of the folks at SiteMeter may have been paved without the benefit of Usability Testing; or at least if SiteMeter did engage in Usability Testing that they overlooked the use cases common to their client’s workflow.

So what has this got to do with my church and/or charity website? Glad you asked …

First, let’s talk about some common pitfalls that I’ve seen ensnare more than one church webaster in upgrading a site:

  • if it ain’t broke, don’t fix it
  • you are not your user
  • even if it is broke, don’t fix it unless you have engaged in some form of usability testing
  • don’t engage in some form of usability testing until you understand how your users consume your system
  • just because someone else does it, doesn’t mean it’s going to work for your demographic

Notice how none of the points have to do with technologies such as Flash, HTML, Ajax, etc … Instead, they all have to do with understanding the objectives of your website – and how your users consume the services your website offers.

And just how do you understand how your users consume your services? Glad you asked …

While one could spend hundreds of thousands on hiring/renting out a Human User Interface lab the likes of Carnegie Mellon and/or University of Maryland – there are also ways of getting “good-enough” results at about 1/1000th of the cost by engaging in some home-grown usability testing that include the following steps:

  1. test 3 or 4 users that range in age, gender, and geekness
  2. setup the tests in a classroom or conference room
  3. provide the user 2 or 3 real-world actions based on services your site provides
  4. only assist the user if they are entirely stuck, otherwise observe how the work around a problem
  5. have both a developer and non-developer in the room, with only the non-developer issuing the test and providing any assistance

The keys to successfully engaging in the above is to:

  • understand what it is you are trying to do with your site (conversion goals)
  • understand who your users are and how they consume your services

Here are some links to some of the other sites not-so-happy with the recent SiteMeter change – listed here to provide a clue as to what type of objections one can meet by pushing out changes without completely understanding the site’s conversion goals and user workflow: