How to quantify downtime

Calculating downtime’s drag on productivity and profit can help make the case for network improvements.

A conservative estimate from Gartner pegs the hourly cost of downtime for computer networks at $42,000, so a company that suffers from worse than average downtime of 175 hours a year can lose more than $7 million per year. But the cost of each outage affects each company differently, so it’s important to know how to calculate the precise financial impact.

By achieving just the average amount of downtime of 87 hours per year, companies could save about $3.6 million annually. And for companies that rely entirely on technology, such as online brokerages, trading platforms and e-commerce sites, hourly downtime risks can be $1 million or more, making availability an even greater concern.

Quantifying the cost of downtime can help you gain funding for technologies that enhance performance and mitigate downtime risks. Yet most organizations have a difficult time calculating the losses associated with downtime because of its complexity.

Sometimes, downtime can cause a loss of productivity for a single user or a workgroup. Other times, the scope is more serious and affects a core application , business process or department, such as a call center or brokerage desk.

Duration is also a critical factor. A loss of a few minutes to an individual or group easily can be made up if employees stay late, but when downtime stretches to hours or days, the loss is more permanent. Whenever downtime impairs business transactions, the length of the outage carries serious consequences. Transactions might be queued automatically during short periods of unavailability, or perhaps clients will call back. But when the event lasts hours, transactions can be invalidated or clients permanently lost.

To quantify downtime there are two primary factors: productivity losses and business losses. Productivity losses affect individual or workgroup productivity, while business losses affect transactions or cause customer losses. Calculating both reveals wasted expenses and lost revenue.

For productivity losses, calculate the downtime based on the effect to users – usually using burdened salary figures. Burdened salary includes user compensation, estimated at $24 per user per hour in the U.S., plus the burden of taxes and benefits, typically 26% or higher than the base salary, according to U.S. Department of Labor. The downtime productivity loss calculation is typically represented as:

•  Number of users affected multiplied by the percent effect on productivity multiplied by the average burdened salary per hour multiplied by the duration of downtime equals downtime impact.

For business applications or groups, the calculations become more difficult. There are two basic methods for the business impact calculation:

•  Number of users affected multiplied by the percent effect on productivity multiplied by the average profit per employee hour multiplied by the duration of downtime equals downtime impact.

•  Number of transactions per hour multiplied by the percent of affected transactions multiplied by the average profit per transaction multiplied by the duration of downtime equals downtime impact.

Consider this real-world example that illustrates the effect of downtime. Accidental changes to Active Directory brought down Internet access for a large financial services firm’s trading desk. As users logged on, they could not access vital information or mission-critical applications.

Some users who had not logged off the previous day were not immediately affected. However, as the policies refreshed, more users became subject to the errant Internet settings.

As a result, the service desk received increasing numbers of calls during the day as the propagation of the unintended change increased.

After eight hours, the problem was traced to an accidental change made by the Active Directory administrator, and the proper settings were restored. The trading desk was affected, and this incident cost the company millions of dollars in productivity and lost business.

To justify best practices, tools or infrastructure that help reduce the risk of snafus that affect availability use a probability equation similar to insurance risk analysis. To predict the effect, estimate the probability that one of the risks will be realized, and estimate how long the downtime will be. The downtime costs can be predicted as:

•  Predicted downtime costs equal probability of event (percent) multiplied by the estimated duration in hours if the event occurs multiplied by the cost per downtime hour.

Once the predicted downtime costs for all the various types of scenarios are estimated, the cost of the people, process and technology improvement to reduce the downtime risk can be compared against the probability and cost of the risks to help justify the solution and assure that benefits can be derived from the assurance investment.