For the most part site owners are beginning to see the value in tracking their sites performance.  Several surveys have indicated that most people gather the data and never do anything with it.  That’s not the best idea as reports are nice but the whole point is to measure improvement.  However, there is a growing number of people that do take action on the numbers.  Unfortunately, many are finding that their actions are negatively effecting site performance.  Take time to critically analyse site data to make sure that the prescribed solution is not an oversimplification of the problem.

I recently encountered a site that had been performing moderately well, providing small niche recorded material.  The site owner had let data compile for two months and sorted through it to see if he could improve orders.  He was reasonably pleased with his traffic but felt that pages were being abandoned too much.  He had a 5% order rate. 

His conclusion was that people weren’t getting the information they needed because the time on page was small.  For his products index page the average time was 2 seconds.  Since it had the lowest amount of time per page he decided to focus his efforts there.  He set out to increase that time and felt some revised content that was more in depth would help conversion.

Here’s the problem, the products index page only listed the items he had for sale with a photo.  It was a good thing that people were only there for a very short period.  Page tracking showed that 93% of them moved off to one of his half dozen products.  The page was working perfectly by getting people to the information they wanted.

He added descriptions to the links and found that time on page increased to around 25 seconds.  He was pleased until he found out his monthly revenue dropped slightly three months in a row afterward.  Upon further analysis we discovered that he had gone from a 93% rate of people making it to the individual product page to a 79% rate.  Site abandonment on this page went from 4% to 7%.  People couldn’t find the product they wanted as easily and were getting lost in added descriptive paragraphs.

The site owners oversimplification of the numbers blinded him to user experience.  He tried to apply a “universal rule” to his numbers and found it was actually detrimental. 

Understand who uses your site and how.  Always think critically about why analytics are showing what they are showing before making changes.  Most importantly track changes you make so that if you miss the mark on an optimization you have the ability to recognize the error and correct it.

Most companies and organizations wouldn’t be content with just having a site.  Many people settle for that, but it isn’t their ideal situation.  Most people want to generate revenue or some kind of action item from their site.  Tableless design does nothing to promote that.  It’s an exercise in clean code and to some extent, accessibility.  That by itself won’t create any conversions on a website.

My rule of thumb is to make tableless or W3C compliant code on the “nice to have” list.  Priority should be placed on meeting business or organizational goals.  Put resources in place to ensure people can locate your site and have clear direction to convert to your desired next step.  If resources are left, then it’s time to explore ideal ways to build the site.  Make sure the end goal is met before moving on to secondary preferences.

I think the biggest problem I currently have with tableless design is that it that it can be a problem displaying on different browsers.  Table design has already gone through these growing pains.  I rarely find an issue anymore with a browser or operating system displaying a tabled design inaccurately.  The browser code has been refined over a decade and a half to make rendering HTML pretty consistent.  Tableless design is way behind the curve here.  Browsers rarely render sites exactly the same and tableless design is subject to all the compatibility problems sites used to encounter 10 years ago.

It is common to see rendering mistakes on tableless sites, even ones built by large specialized design firms.  In many cases, it doesn’t bother me. If a picture bumps to the left a few extra pixels, who cares?  The problem is that it can often render so poorly that text gets cut off or overlapped.  That’s a major problem.  That is a problem that can negatively effect the user experience and site conversion. 

Having said all that, designers and developers are getting more savvy at tackling these problems.  The trouble is that it usually requires (what my last post warned of) a stock layout with little variation from other sites or additional coding to render more accurately across platforms and browsers.  Those might not be deal breakers but something to consider.  Extra cose is also a common explanation for increased setup fees.  Usually the problem is that no one can foresee all of these issues.  Even the best designers and developers that test for rendering problems are likely to miss some.  The good news is that the testing can minimize risk of a bad user experience.   The bad news is that some visitors will likely encounter a problem.  I’m not sold that WCC compliance with separate CSS layouts is worth that.  There are few things that compare to a good user experience and site conversion. For me, tableless design is not one of them and often looks like an unnecessary risk.

I am by no means an expert on layouts with tableless design, so this is based on my perspective of the sites I’ve witnessed being produced.  They all have a similar layout.  There is a banner along the top, often a set of links along the right or left, a footer, and a blank area in the middle with content.  The imagery and colors change but the fundamental structure is static.  While there isn’t anything wrong with that, it is limiting.  There also seems to be a standard set of roll over options that tend to pop up.  I suspect this is because many designers are reusing others code in an attempt to be “modern”.  If tableless design does limit layout, it can make a site seem  less professional because of a “template feel”.

I tend to believe that this isn’t actually a problem with the technology but with many designers/developers.  While the structure of the layout seems pretty static (if anyone knows of a unique tableless layout with unconventional or custom layout, please leave it in the comments)  I have to believe that given enough time and expertise that almost any layout could be achieved.  The problem is that many designers/developers want to be tableless and WCC compliant but don’t have the knowledge to build a site like that from scratch.  So they end up using preexisting templates and claiming them as there own.  This is extremely common when a separate blog is created.

Why is this a problem?  The first is that you aren’t getting what you pay for.  Building small customizations on top of a preexisting template is not that challenging.  Furthermore, unless significant testing is done, it might cause layout conflicts down the road.  Finally it makes the site more generic.  Templates have a way of surfacing in a lot of places and it always detracts from a site if people see the layout as a cookie cutter of someone else’s.  Most company’s are carbon copies of their competitor so their site shouldn’t be either.  Especially if they’ve bought a product they believe is custom.  While table design does have draw backs, I’ve seen just about any layout succeed.  The method has had years of innovation where designers and developers can achieve some extremely impressive layouts on the web.

So if you want a tableless design, make sure you understand how the layout will be created and that there might be some restraints on the look.  Is it a custom site or is it a template with a few changes made? Neither choice is necessarily wrong, but you need to understand the pros and cons and should be seeing a discount on any template tweaked site.