For as long as shopping on the internet has been possible retailers have been battling the ‘Big Days’ in their calendars. Tackling the problem of how to handle vast spikes in traffic for short periods of time, especially when those spikes mean revenue, is neither trivial nor cheap. Thus from the early days of Interflora using Intershop to drive online sales for Valentine’s Day in the late nineties, through to the growth of online Black Friday and Cyber Monday in the late noughties, the same problem has reoccurred; the website has been overwhelmed by users.
In 2007 Black Friday was still a US only event. E-commerce was at a pre-crash peak and Americans had full credit cards to use. Amazon was pretty much the only major retailer not to go down that day. Since then (with full credit to Amazon) the Cloud has arrived, bringing with it scalability and flexibility – the very things that were hampering the ability of retailers to respond to rather than engineer for peaks in demand.
At the same time the growth in the market has been staggering, year on year Black Friday online sales growth from 2012 to 2013 was still 34% for mobile and 15% overall. That’s major double digit growth for established retailers like Target or Walmart. Growth they can’t and won’t replicate in any other areas of their business. So there is huge demand, multiplied by the power of the internet to globalise markets. Black Friday is now not just a ‘thing’ in the US, it is a feature of retail calendars from Copenhagen to Nairobi.
But, despite the opportunity and the criticality of events like Black Friday and Cyber Monday, numerous websites keep falling over and it’s not just small retailers. In 2013, the Black Friday casualties included the likes of Walmart, Motorola and RBS. To underline that problem the tech press is awash with “thought leadership” and “comment” articles by website monitoring and load testing companies, indirectly plying their wares, albeit perhaps a bit late.
The downside risk of your site falling over is huge, recent research by Akami showed that 64% of online shoppers disappointed by the experience will shop elsewhere next time. That’s 64% of your sales last year; plus 64% of the 15% your sales grew. If as the CTO/CIO of a retailer you are prepared to risk losing 74% of your customers by letting the website go down you should be polishing your CV. It is an utterly unacceptable FAIL.
There is plenty you can do: Plan; understand the business requirements and the demand you may face analytically and not based on a recon. Test everything repeatedly; if possible continuously, then test it again a couple more time. Develop your apps to be agile and scalable. Mitigate the impact of spikes; put a queue in to your system to control the flow of customers or perhaps only take payment when the goods ship. Use the cloud; but make sure your code and your delivery pipeline is as good as your cloud vendor. And don’t forget, even in e-retail, it’s an omni-channel world.