Web performance – Are you a smooth operator?

Your website may look the business, but it’s no good if technology behind the scenes lets it down. SciVisum chief executive Deri Jones reports


Everybody buys everything on the web these days. Well, not quite true, but when my mother (aged 75) tried to use eBay (albeit with the help of neighbours), it seemed to me that such a bold statement is more than half right.


But not everything is rosy in the travel garden. It turns out that the Brits who buy holidays online are also the least tolerant of poorly performing websites. Some 16% would never use a website again following poor performance, and a quarter would turn to a competitor, or the high street, if they experienced problems on more than one occasion.


SciVisum’s recent study confirms the huge commercial value of holiday shoppers to online companies and the risk of financial loss if they fail to guarantee smooth transactions on their sites. What we also discovered was that consumers who buy holidays online shop more frequently and spend more than average online shoppers. Good news for readers of Travolution!


On average those who buy holidays spend around £118 a month compared with an average monthly spend of £88 by all other online shoppers. More than one third shop monthly (36% versus 24%) and a quarter do so once a week, compared with 15% of other online shoppers.


They also spend more on a single purchase, with a huge 96% investing £100 or more on a single item.


Interestingly, the battle of the sexes doesn’t seem to have reached the online travel sector, 63% of women versus 57% of men buy holidays, and across the UK, central London has the largest percentage of consumers booking a break via the Internet.


Those living in the south, plus London and Scotland are also more likely to turn their backs on high-street travel agents than their counterparts in the North and Midlands.


For shoppers choosing to book their break via the Internet, we found that speed and price are the motivators for going online. Some 66% gave both as their main reasons for doing so while only 57% of other shoppers claim to be driven by the same factors. Other influences cited by online consumers include a wider range of choice and the avoidance of high-street queues.


Nevertheless, while consumers who buy holidays online regularly spend more, listen up e-commerce managers: they are also the shoppers most prone to web rage.


While more than a quarter (27%) would forgive a favourite website up to twice for poor performance, 86% would leave a website with an incomplete transaction, compared with 78% of all other online shoppers.


For well over half of online shoppers (64%) the biggest indicator of poor performance is an online shopping basket crashing, followed closely by being unable to amend an order (60%). This becomes double jeopardy for e-travel managers when combined with another uncomfortable e-commerce fact: travel sites do more web-farm work per visitor than other e-tail sectors. So not only are the consumers more picky, but the sites are more prone to performance problems – certainly in the 24/7 testing of user journey effectiveness we do for clients, the travel sector has the worst performing sites on average.


Why do travel sites’ web farms have to work harder? Sometimes because travel organisations are not geared up for real-time stock control. Maybe internal mainframe issues mean nightly or four-times-daily updates are passed to the web farm, which then has to bolt on top the ability to keep track of what it has sold since the update.


On top of that, online stores worry about shoppers searching and finding a package they like, but it being sold out by the time it takes them to complete the five to 30-minute add-to-basket and check out journey. It can happen at busy times, especially if the number of places left gets down to single digits.


On that basis, many travel sites do clever behind the scenes ‘temporary reservations’, i.e. as soon as the shopper clicked on one of the offered holidays from their search, that package was removed from the database, so that another shopper would not be able to see or buy it. However, this all means extra work per visitor for the web farm and not just extra server load, but more scope for new problems to be introduced. 


Another common feature of travel sites is the clever caching code the software developers write, hoping to reduce the server load by having search results saved in a cache, ready for quick reuse for similar searches that occur within the following five, 10 or 30 minutes. The impact of that caching, if not very carefully implemented and maintained, can mean that future small changes to the site can increase user frustration, because holidays selected from searches, turn out a few pages later in the journey to not be available.


On travel sites, when testing the vital money-making user journeys on add-to-basket and check-out we see commonly that journeys fail 1% or even 5% of the time, simply due to these behind-the-scenes web farm problems.


These failures are a shock to e-business folk, and invisible to the technology people whose logs only show errors that the web server itself spotted. This excludes the ones where there’s ‘nothing broke’ in the technology, but the user experience was certainly far from perfect.


As E-consultancy.com advised in October, it’s time for e-marketers to take back control of web technology; and that means you need to be measuring your money-making user journeys around the clock.


If someone suggests that simple uptime/downtime monitoring of your homepage and/or a few main pages will suffice, just ask yourself if it’s the homepage that makes the money? Not at all – it is the whole multi-page user journey, from finding the package, adding it to a basket and checking out.


If you can set up a regime where the performance of those user journeys is used as the internal KPI between technology, marketing and customer service teams, then you’re on the road to increased sales. This is one situation where the old chestnut is spot on: if you can’t measure it, you can’t manage it.


Once you’re measuring your key journeys, it’ll soon become apparent where the problems in the technology have the most impact on shoppers. You’ll find out which pages in which journeys are the slowest on average, or more usefully, which pages are not so bad on average, but when they are slow they are really slow – maybe five or 10 or 20 times as slow as the average.


This information is then gold dust for the tech teams – it allows them to apply their skills in a very focused way to just the pages and the parts of the web coding that are impacting visitors.


They’ll get much better bang for their buck from this, than trying the ‘make the whole website faster’ approach.


Deri Jones is chief executive of SciVisum, user journey-based web effectiveness testing specialists. SciVisum’s services measure the performance and functionality of clients’ business-critical online systems.

This website uses cookies to ensure you get the best experience. Learn more