E-retail web site errors causing big hidden losses

WM Circle Logo

By: Wise Marketer Staff |

Posted on November 2, 2007

E-retail web site errors causing big hidden losses

British e-commerce web sites are thought to be the cause more than £300 million worth of sales each year, and they risk driving customers to competitors because of 'invisible errors' that can't easily be detected by standard web analytics, according to online performance consultancy SciVisum.

One in three (33%) of consumer-facing 'online journeys' tested by SciVisum experienced more than 3% error rates, while more than 10% demonstrated extreme inconsistencies in the delivery speed of the overall journey.

Trouble for Christmas "This is a worrying trend for e-commerce and IT directors, as well as for consumers, with Christmas just around the corner. Poor performance and unnoticed web site errors will mean lost sales," warned Deri Jones, CEO for SciVisum. "Companies must check the performance of their sites in advance, and those that fail to do so might as well include a link to their competitors' sites."

The latest SciVisum Lost Online Sales Study, which investigated the performance of 40 e-commerce web sites from the retail, finance, insurance and travel sectors during the past six months, found that customers of e-retailers are being exposed to a significant number and range of problems that prevent them from completing their desired journeys (i.e. shopping online).

Invisible errors Invisible errors are not outages affecting 100% of web site users, the company reports. They are worse in many ways, being problems that impact only a percentage of users at any given time. A problem that affects as few as 1 in 100 users is not always easily reproduced by IT teams, so these issues often remain unresolved.

SciVisum's testing adopts a mystery shopper approach adapted specifically for the world of online shopping. A representative actually visits each site being tested and tries to make a typical user journey every five minutes throughout the day. This allows the company to see what customers see, and makes it possible to find out if there are any intermittent problems that affect real users while remaining invisible to other analysis.

Most common problems The most common problems the company reports encountering include:

  • Session swapping, where two users see each others' online sessions. Nowhere is this category of problem detected in server or analytical logs.  
  • Page not delivered: Because the web page is not actually delivered, there is no log of the error in web analytics.  
  • Jump back: the user is in error forced back several pages. The new page is itself a valid page, so no errors logged in analytics or tech logging.  
  • Page content incomplete. This is a serious problem, as web analytics will show that a page was delivered, but not whether the user actually received the whole page in their web browser.  
  • Shopping basket errors. In The Wise Marketer's experience, this is an alarmingly common problem, in which the user finds their shopping basket is empty or incomplete after adding items. Often caused by incorrect handling of web cookies (by the user's web browser) or poor programming technique, this category of problem is not detectable in standard web server or analytics logs.

The need for speed The research also highlighted inconsistencies in the delivery speeds of the journeys that users undertake on e-commerce web sites. More than 30% of journeys experienced performance varying by more than 200%, with 10% varying by more than 300%.

Because different technologies are used to service the different routes that customers follow, those journeys can often provide very different experiences, even though they run on the same web site. Variations in performance means that returning visitors will be frustrated on the occasions when the web site runs slower than expected, and first time visitors are likely to be driven to competitors because they (quite naturally) assume that the web site always operates at that speed.

Peak load problems Frustratingly for consumers, web site performance tends to be worst in the evenings (between 8pm and 10pm). This is also, by no coincidence, the usual time of peak internet traffic levels. But it all adds up to web sites performing at their worst when it will inconvenience the most visitors.

But this problem is also often invisible to the e-commerce manager, because simplistic measures (e.g. page speed average per day) hide the fact that just one or two of the core user journeys perform really badly for a few hours each day.

Deaf ears in the works According to Jones, "The UK's online landscape is plagued by these errors and as users continue to become web savvy and increasing numbers of people encounter them, they won't remain invisible for long. Interestingly, we often find that our clients call centre staff are aware that there are problems that are invisible to colleagues in other departments. It seems that a deaf ear is turned to their feedback; being repeatedly told by IT that 'no problem was found' or that 'the problem couldn't be reproduced'. That last phrase is a typical internal response to the kind of sporadic, invisible errors we found, and means that the underlying problems will never be addressed."

Based on the findings, SciVisum suggests that e-retailers should adopt a mystery shopper approach and test the paths that real users take when making use of the web site. Simple uptime and downtime monitoring of a few key pages is not sufficient to find these hidden problems.

At the same time, the company suggests a stronger focus on relevant performance data. To help find otherwise invisible errors and performance bottlenecks, companies will need to focus on using data gathered by simulating real user journeys and experiences, rather than data from internal servers or monitoring services.

But most importantly, management must take full ownership of the problem, and financial, strategic, and marketing personnel need to work with the IT department and accept only metrics and data that is relevant to a real end user's experience.

More Info: