Valentine’s Day: the Worst Day to Break Up with Your Customers
by Alex Painter, on 2/21/18
Love it or hate it, Valentine’s Day is big business.
It’s hard to avoid the ads, the promotions and the special offers, all dedicated to getting us to part with our hard-earned cash in return for some romantic gesture or other.
Away from the high-street stores, the restaurants and the garage forecourts, many consumers were doing their Valentine’s Day shopping online
We decided to monitor ten of the more popular florists’ websites to get an idea of how well they had geared up for one of their busiest days of the year. In particular, we were looking for evidence of how a peak in traffic might have affected load times and availability.
Our experience of Black Friday last year led us to believe that most would do relatively well, with retailers getting ever more sophisticated when it comes to preparing for peaks.
Sure enough, most of the sites we looked at had no major issues.
The chart below shows the performance of one of the sites we tested between 10 and 14 February. Successful tests are shown in green, with load times represented on the y axis (the grey area indicates page size). This is exactly what they would have wanted to see during peak – almost perfectly consistent load times throughout and no outages.
The worst time for a site to be down
However, it looks like one florist fell foul of a surge in last-minute panic buying on 13 February, suffering an outage between 16:15 and 16:45.
In the chart below, the outage is represented by black bars, which in this case indicate that 503 responses were returned (the surrounding bars are yellow rather than green because of an unrelated issue with a third-party resource):
Happily, it looks as though the site was up and running again relatively quickly. But it’s almost certain to have been a costly experience for the retailer. Frustrated would-be customers had plenty of other options, delivering an unexpected windfall to competing sites.
The importance of preparation
This is a timely reminder of the need to test in advance of peaks to understand how a website is likely to perform under pressure.
We have no way of knowing whether this particular retailer failed to load test its site or whether it did test and just saw far more traffic than predicted. Peak periods also present a dilemma to retailers who might think twice before investing in scaling up just to cope with a surge in visitor numbers one or two days a year. Perhaps this retailer was well aware of the risk but lacked the resources to prepare adequately for it.
However, outages during peak periods are particularly damaging, not just to short-term revenue, but also to reputation. If you fail to show enough love to your website and your customers, you’re in danger of losing both.