Amazon.com, Google and Nasdaq all had high-profile — and costly — website outages in August. This most recent outage for Amazon lasted 30 minutes and potentially trimmed nearly $2 million off its bottom line. Google's five-minute blackout in mid-August cost it $545,000 in lost revenue. And Nasdaq's three-hour hiatus froze trading and disrupted other markets as well.
Retailers typically focus on improving the online customer experience through personalization, omnichannel integration, rewards programs and other incentives, but as Amazon, Google and Nasdaq demonstrated, none of the bells and whistles matter if your site is unavailable.
Consumers today expect 100 percent availability, so during and after any outage, organizations typically have two questions top of mind: One, what caused the outage and two, what did the outage cost?
We don't always learn the cause of these high-profile outages, and if we do, it's often not until after the fact. However, we don't have to wait for root cause analysis every time a website is offline since the cause can usually be traced to only four areas, which are also the root of the total cost of ownership for any site. Here are the four areas as well as the costs that every retailer needs to keep in mind to avoid downtime:
1. Research and development (R&D): This is the cost associated with having software developers implement functionality of the system. In the case of a website, the effort is spent on customization of systems like ATG, databases, adding drop-ship functionality, tying rewards to purchase history, and so on. The more complete this effort is, the less likely you'll experience a costly incident.
2. Testing: Let's face it, no research and development effort is error free, so the more investment there is in testing before production, the more likely it is that errors are caught in advance and the lower the downtime risk.
3. Hardware: The type and quality of hardware has a significant impact on website uptime. Imagine a large retailer using old routers, load balancers and servers that are prone to failure and can't withstand seasonal shopping peaks. High-quality hardware and a solid configuration lower the chances of an incident.
4. Operations: Infrastructure management, the cost of software licensing and the cost of labor all add up. For example, the quality of the data center where the retailer's system resides has a direct link to site uptime. So too does the cost of labor. A new grad, for example, has a different cost than an expert system administrator or database administrator who has a higher level of skill and reliability. Keep in mind however that a higher cost of operations typically lowers the likelihood of an incident.
And, of course, the final cost in the total cost of ownership equation is the cost of any incidents. As we saw in August, even a few minutes of downtime can erase hundreds of thousands of dollars in revenue.
It's not enough to simply provide a better experience for shoppers. As this summer's website outages have shown, companies need to think first of the fundamental areas supporting their system. And while R&D, proper testing, up-to-date hardware, and operations costs can be steep, they're likely far below the potential cost of a downtime incident. By focusing on these four areas of your system, you'll be able to make sure your online shop remains open for business.
Samer Forzley is the vice president of marketing at Pythian, a global data management consulting company that specializes in planning, deploying and managing mission-critical data infrastructures.
- Companies:
- Amazon.com