It was the height of the Pokémon Go sensation, way back in the summer of 2016. YouTube celebrities were born overnight. Millions of kids emerged to play the augmented reality game and feel the first touch of sunlight on their skin in months. Kids were actually getting exercise again, and Wall Street was making money.
My 9 and 6-year-old boys were not immune. We joined the legions. I took them to parks where the PokéStops and PokéGyms were located and mingled with neighbors out with their kids playing the game. My boys fell in love with SlowBro, a somewhat rare Pokémon I was to find out. So I did what any dad would do. I went online to find and purchase a plush toy of the Pokémon. But, I was in for a shock. Never was it so difficult, ever, to buy what I wanted. And, here’s why.
It was late July when I tried to “catch” the elusive, stuffed SlowBro, about 2 weeks after Pokémon Go exploded with 100 million downloads to mobile phones almost overnight. This meant all those Pokémon ecommerce sites, averaging a few hundred inquiries a day back in June 2016, were suddenly hit by a million inquiries an hour. The impact on my customer experience? Click, spin, close tab, next. The result to the webstore server? Sputter, spark, smoke.
Not Just About Infrastructure
In Wolfgang Digital’s Ecommerce KPI Benchmarks Study for 2017, the average page-load time for ecommerce sites is 6 seconds. This is 3x the recommended 2 seconds or less. And when 40% of people abandon a website that takes more than 3 seconds to load, you can begin to see how an ecommerce store or sales channel without the proper infrastructure to handle capacity can lose to the faster competition, and even go bust. For millions of online shoppers, if it’s “click and spin” for more than three, long, impatient breaths, it’s on to the next site.
You might be thinking to yourself – hey, couldn’t the cloud, with their elastic load-balancers distributing traffic across a “cloud” of servers, have helped these Pokémon specialty sites? The answer – most of these ecommerce sites were already hosted on cloud-based domain servers due to the restrictive costs of owning and maintaining a server on premises. So, apparently not. What went wrong?
Even in the best enterprise cloud environments, such as AWS, GCP, or Azure, the load-balancing and auto-scale only goes so far. When a site normally gets a few hundred inquiries a day, and that suddenly spikes to a million an hour, the automated load balancers are not going to keep up, and your site is going to go down. That’s when the alarms go off, and people scramble to get the server back up – provided they have the people or have paid for the service. Most of those Pokémon webstores were on economically priced domain cloud servers with architectures for small businesses, not Enterprise Cloud architectures like the big 3 mentioned above.
Working with an expert managed application service provider to pre-scale capacity in anticipation of seasonal spikes or promotions is a smart thing to do if you want to increase conversions and fully capture the potential revenue. That same team will also be there to quickly handle events where no amount of pre-scaling or auto-scaling can anticipate the flood of inquiries, as was the case of the perfect Pokémon storm.
With Ecommerce, Speed Matters
The ecommerce sites that have survived over the years have learned this, often the hard way. The same Wolfgang Digital study showed that in a world of 6-second page load averages, reducing the load times by just 1.6 seconds increases annual revenue growth by 10%. In the case of those Pokémon ecommerce sites, in that 1 month, it probably would have been closer to 100% or 200%.
I eventually bought the SlowBro plush toy a couple of days later from a large retailer site. Picking up on the latest trend, and quickly bringing about their considerable resources to capitalize on it, the retailer sourced the plush toys and quickly got the products online. My transaction took less than a minute.
I sometimes think about the millions of Pokémon dollars that were made during that time. Those windfall dollars passed right by those small and mid-sized ecommerce businesses, some of which are no longer around. These were transformational dollars for companies that could have overnight brought new jobs and new dollars to a community that probably needs it badly. Instead, those dollars were sucked in by large online retailers just to add a little blip to their quarterly revenues of billions. Those are the consequences when short-sighted budget concerns win out over innovation. That’s how to go out of business in 6 seconds, or more and why speed matters!
Latest posts by Steve Sassi (see all)
- Speed Matters – How to Go Out of Business in 6 Seconds, or More - November 11, 2017
- Get Your PIM Off the Ground (And Into the Cloud) - October 11, 2017