During Amazon.com’s much-anticipated Prime Day sales, customers experienced significant outages, affecting sectors across the brand’s retail outlet and even extending to its Amazon Alexa products in home. This resulted in trending #AmazonPrimeFail and #Amazonoutage across social media from outraged consumers, potentially significant revenue loss and many of us asking how this can still happen in 2018.
Earlier this year, we witnessed the economic impact felt when major retailers close business for even a few hours. In April, Starbucks closed 8,000 of its 27,339 global stores for a single afternoon. The racial bias employee training initiative is estimated to have cost the company about $12 million in lost revenue, as calculated by Marketwatch. Such mass closures are rarely seen with brick and mortar stores, but it helps put into perspective how detrimental it can be when customers are unable to access their go-to brands.
Compare that with online outlets, like Amazon and its Web Services, which thousands of other businesses rely on globally. When this single platform faces a glitch, every business along the chain suffers revenue loss instantly.
A 2017 outage at AWS caused by a “typo” in a command executed by a team member within their S3 billing system cost companies in the S&P 500 index “$150 million, according to Cyence Inc., a startup that specializes in estimating cyber-risks. Apica Inc., a website-monitoring company, said 54 of the internet’s top 100 retailers saw website performance slow by 20% or more.”
The sheer volume of transactions that occur in major online retailers reliant on AWS each and every second means the economic impact that would result from such a glitch can add up very, very quickly.
Amazon is such a powerful force in today’s e-commerce sector that its net revenue is larger than the GDP of several countries (like Qatar and Algeria). In fact, based on its annual net revenue in 2017 Amazon would be the 55th largest country in the world.
This week, just as Amazon.com’s much-anticipated Prime Day kicked off, featuring the best deals of the year across the site, customers in certain regions experienced outages for close to 13 minutes, and continued to experience outages throughout the day. Some estimate this could result in over $2.6M of revenue loss.
Let’s look at how the impact of an outage like this one trickles down.
Brands and manufacturers invest major sums into Amazon Prime days and marketing their products on the site year-round – on the low end, upwards of $750 per product to advertise during a four-hour window, and tens or hundreds of thousands of dollars for exclusive placements and deals on the high end. As a result, there’s a lot riding on customer access to that product during those precious minutes.
As well there is Amazon’s massive global logistics and distribution chain, countless divisions of teams within the organization, not to mention third party agencies who exist solely to help brands crack the Amazon puzzle. During an outage like the one experienced this week, all of these parties are rendered helpless and ostensibly frozen, bleeding money every minute the website malfunctions.
This sure makes you appreciate your web dev team.
Economically, this is not a laughing matter. This isn’t the first time Amazon has seen such an issue, but in 2018 we hardly expect it to keep happening, particularly on a day as important as Prime Day.
This one extended beyond the shopping portal, into homes, where Amazon’s Alexa smart speaker was momentarily facing glitches and unable to answer requests.
To make matters worse, in Europe, members of Amazon staff took the opportunity when the world was focused on the retailer to strike against its working conditions like proposed cuts in wages and health benefits. According to WIRED, “nearly 1,800 workers went on strike on Monday in Spain, where the planned protest was first conceived as a way to fight pay cuts and restrictions on time off.” They were joined by workers in Poland, Germany, Italy, France, and England.
What it takes to keep the world’s leading Cloud service running
Looking at AWS Cloud overall, AWS Vice President and Distinguished Engineer James Hamilton noted at re:Invent 2016 that, “each day, the company adds server capacity to its network that is equivalent to what it had at the end of 2005, when it was an $8.5 billion company.” It organizes its data center infrastructure into 18 regions, across 55 Availability Zones, each of which has from one to as many as eight data centers – and more on the way. Data centers range from 50,000 to 80,000 servers. (For anyone keeping count, that’s somewhere between 2.75M and 35.2M servers.)
In fact, AWS uses a forecasting model driven by machine-learning research to make decisions about when to purchase more server space, and where, as well as where it can store excess components in the event that it needs them. Nevertheless, the giant is in no way immune to technical glitches and malfunctions.
Despite the outage crisis, boycotts and negativity online, US Prime Day sales outperformed last year’s: online shoppers spent 54 percent more in the first three hours of Prime Day this year, than the first three hours last year, according to a report from Bloomberg.
It was also a good day for CEO Jeff Bezos, who just became the richest man in modern history, with net worth rising above $150 billion on Monday. According to Bloomberg, that’s about $55 billion more than the world’s second-richest person, Microsoft co-founder Bill Gates.
Time will tell whether the event will have an impact on future Amazon Prime Day sales, and whether brands investing heavily in these events will begin to look to safer outlets to ensure consumers are able to access their products. But interesting to note is just how quickly consumers can turn on an otherwise tried and true favorite in moments of inconvenience and frustration.