Netflix needs metrics for reliable delivery
My old college business professor would love Netflix. He could build a whole semester’s worth of subject matter studying Netflix, alone. From innovative service model to adaptive positioning techniques, Netflix offers a wealth of case studies for students. Today’s lesson: the importance of learning from past mistakes.
Netflix’s original business model was based on delivery of DVD movies via the postal service to a subscriber’s home. By using a software model that is able to manage a customer’s movie queue and harnessing the United States Postal Service’s existing distribution channel, Netflix could deliver a movie to your door in a day or two. Back in March of 2008, Netflix experienced a “glitch” in its delivery service software and movie shipment was delayed for 24 hours. Netflix credited customer accounts for 5% of their monthly viewing bill in reparation. When a business suffers an interruption in service that is company wide, there are usually studies and action plans that result in the implementation of corrective measures to assure that the same problem doesn’t resurface. Netflix is, apparently, an unusual company.
For a period of three days, Netflix, again, experienced a significant failure of its shipping system. This impacted a sizeable chunk of the 8.4 million members who subscribe to Netflix for a steady flow of DVDs via mail. From Tuesday through Thursday, only partial shipments were able to be made due to a “pretty severe technical” issue, as Steve Swasey, representative for Netflix, told the Wall Street Journal. Netflix web delivery of movies was unaffected. And while the crisis appears to be over, managers and business students everywhere should be asking “How did this happen again?”
If a company is to maintain reliable service, it must have a means of measurements in place to allow for failure prediction. Very rarely will a computer, or even a mechanical process, suffer catastrophic failure. Over time, links in the process deteriorate and slow down to substandard levels. For mechanical products, regular maintenance and cleaning will keep a machine in top form. For computer systems, monitoring software is available with threshold alarm functions. This allows for predictive monitoring to be put in place and used effectively. Predictive monitoring gives businesses an opportunity to catch and address an issue as a slowdown, well before it becomes a failure. While the initial cost for such a package may seem high, the costs in lost production, loss of reputation, and in customer refunds are much higher.
There are quite a few things Netflix has done right. As covered in previous Talkibie articles, Netflix has done a fabulous job integrating its service using new technological delivery methods. And it’s done a great job in customer relations, owning up to failures and taking proportionate steps to win back loyalty. March’s day of lost service garnered a 5% refund to customers. This most recent outage of three days will net existing customers a 15% refund and new customers a full week of free service.
But in the area of learning from past mistakes, Netflix seems to have dropped the ball. Investors and shareholders will want to know what proactive steps Netflix is taking to ensure that its service doesn’t fall into the same trap a third time. The answer is to use technology to keep the company online. Class dismissed.
By Alicen Hogan