Reflections on the Need for Speed
According to the Aberdeen Group, every eCommerce Manager in the land would gladly take a 7% increase in conversion, an 11% increase in page views and a 16% increase in Customer Satisfaction.* These are metrics that online retailers already spend much time and effort optimising. So imagine how frustrating it can be when a whole team and its efforts are derailed by literally 1 second of additional page load time. That one second can completely reverse the desired metrics above.
Worse yet, there may be a multiplier effect when these delays occur. Consider first-time visitors, or visitors to pure play retailers, whose website presents the only means to cultivate a relationship with a customer. Poor performance might inadvertently lead to a higher rate of call centre conversations, further increasing negative impact. As discussed previously in Retail Technology Review, the online shopper’s need for speed can make or break businesses. Speed matters, and it always will.
Here in the UK, the government wants everyone online by 2012, so we are all being asked to pay up for faster and more robust broadband services. As a consumer, I expect fast page load times, and if not achieved, will shop elsewhere. This is fairly typical behaviour: in fact 40% of shoppers abandon sites based on poor performance. As larger retailers continue to set the bar for performance, consumers’ expectations are being conditioned, such that any site deviating from these standards will recognise the inverse of the statistics at the start of this post.
Understanding the gravity of page load delays should give retailers pause: perhaps adding more and more functionality and richer content to a site can be detrimental as opposed to ‘an enriched user experience.’ Each new site enhancement and its role in the bigger picture require careful performance benchmarking and resilience testing to understand its ‘real’ impact.
A solution could be benchmarking against competitors and the perceived ‘best’ sites within a vertical/demographic range – measuring both page load times and 3rd party applications over periods of time. As a retailer, if you were to improve overall performance from 5.7 seconds average page load time to 5.2 seconds, you would feel delighted. But knowing your closest competitor’s pages load in 4 seconds, your perception would be different.
This indeed is the significant point of RichRelevance’s investment in its own IT infrastructure. As a partner of many Tier 1 retailers’ eco-systems we must benchmark ourselves and so utilising a global data centre ‘cloud’ environment built on state of the art solid state disks (SSDs) we can not only justify fast response times (average 60 millisecond response time) but also effective utilisation measurements (typically not surpassing a 10% utilisation of capacity.)
Two final points that are also very pertinent to performance would be internationalisation and mobile. Many retailers are now investing in European growth, so each new venture needs to welcome its additional foreign customers with the same performance as the home site. Mobile users expect even more – they are happy to sacrifice features for faster load times; in this case, Gomez found that 3 seconds wait time seems to be the perseverance level.
For all merchants out there we can safely say that speed is a critical metric to consider and one in which ‘the norm’ will only accelerate as more robust clouds appear. Time is currency—not only for its real impact on a retailer’s bottom line, but also for its undeniable influence on customer lifetime value and brand perception.