How the world’s biggest shopping festival relies on Alibaba Cloud
$38 billion worth in total, across 200,000 brands, and in 78 countries.[1] The record it broke was $30 billion from the previous year. The event run by online retail giant, Alibaba now dwarfs any other sales event -- even Black Friday, with its $6.2 billion in total sales, barely registers a blip in comparison.
It's hard to truly grasp the numbers involved with Double 11 -- they're just that significant. 500 million users make purchases on the day. As many as 540,000 transactions are processed per second. It took just 68 seconds for the sales to top $1 billion.[2]
These numbers are all the more impressive when you consider that Alibaba only started the shopping festival in 2009.
Managing the world's biggest sale
While Double 11 produces some truly astounding numbers and is a major driver of Alibaba's ongoing success as a global leading retailer, the technological challenge of powering this unprecedented sale should not be discounted.
A second of downtime would impact hundreds of thousands of transactions, resulting in millions of dollars lost. That's not even considering how a poor shopping experience can push away future customers and inhibit Alibaba's future growth. For Double 11, Alibaba's cloud environment needs to be able to scale to an extreme degree to have zero downtime while also maintaining the security for millions of people's transaction details.
What's more, the cloud was tapped to offer processor-heavy features to help grow the prominence of Double 11 worldwide. For example, Alibaba's cloud powers a machine translation service on the cross-border AliExpress, and during the shopping festival, this was called on to translate over 200 billion words into different languages.[3]
Scaling for an event of this magnitude requires much more than simply adding temporary capacity to an environment. Alibaba Cloud leverages what it calls X-Dragon Architecture -- now into its third generation -- as a way to meet the challenging demands of an online retailer that is the size of Alibaba. X-Dragon Architecture seamlessly integrates Elastic Compute Service bare metal servers and virtual machines into the one environment and then layers it with rich AI processes that have been designed to increase the number of queries per second processed to decrease latency.[4]
Strategically planning for sales
Sales are the single biggest challenge for retailers with an online presence. Last year, over the Thanksgiving sales period, Costco's website went offline for around 16 hours, costing the company $11 million in sales. Adding further injury to the company, it also faced a social media fallout from angry customers.[5] According to research, consumers are four times more likely to stop supporting a brand after a bad experience -- a website being offline during an important sales period certainly qualifies as that.[6]
Even the biggest e-commerce websites are not immune to outages. Amazon experienced a temporary outage that lasted 13 minutes last year, with the estimated cost to the company, during a non-sales period, amounting to over $2.5 million.[7]
As mentioned, retailers need to ensure that their online environment is robust, available, and highly scalable. It must also have the smarts to be able to scale up and down in real-time to meet with demand.
CIOs should view these five factors as being critical to the health of their e-commerce site, in terms of both uptime and customer satisfaction:
- Availability -- the technology used by the cloud environment should be robust and have full redundancy to ensure 100% uptime. Remember, even a few minutes of downtime, as 99.9% uptime allows, is going to cost a retailer a lot of money.
- Cloud-native -- the environment needs to support hybrid and multi-cloud environments so that it can run across multiple geographic locations and ensure a consistent experience for all customers.
- Scalable and distributed -- this is critical to the ability for the retailer to meet demand during peak sales periods. The cloud environment needs a distribution and replication strategy that is responsive and can process an unusual number of transactions concurrently.
- Performance -- keeping latency to a minimum is critical. Not only because the consumer has no patience for slow transactions, but also because an efficient, high-performance cloud environment minimises any impacts on the environment and saves on the cost of processing transactions.
- Security -- the cloud environment needs to hold credit card data and be in compliance with international standards, such as Europe's GDPR. The potential cost of non-compliance, above and beyond the reputational fallout, means that you need a reliable cloud service that you can trust.
The Alibaba Cloud has proven with Double 11 that it is the only environment with the scalability and stability that can deliver to the most extreme spikes in retail demand. Read more information on how you can run your own retail environment on the Alibaba Cloud and deliver a superior experience for your customers.
[5] https://www.businessinsider.com.au/costco-website-outage-costs-millions-in-sales-2019-11?r=US&IR=T
[7] https://www.upguard.com/blog/the-cost-of-downtime-at-the-worlds-biggest-online-retailer