How Much Does Downtime Really Cost?
Why is it difficult to justify an investment in business resilience when the risk and costs of downtime are so high? One reason is because the current business environment insists on clearly demonstrated financial returns on investment. The other reason is because the risks and costs are not well understood.
Despite advances in infrastructure robustness, occasional hardware, software and database downtime is unavoidable. In fact, Dunn & Bradstreet reports that 49 percent of Fortune 500 companies experience at least 1.6 hours of downtime per week. That translates into more than 80 hours annually. How would your customers react to an 80-hour outage?
Understanding Downtime Costs
There are options for reducing total downtime, in some cases almost to zero, but none of them are free. Hence, organizations must determine whether these options are worth the cost given the financial impact of downtime. Put differently, each organization needs to decide what losses it is willing to accept due to outages and design a solution to avoid this worst case. Average downtime costs vary considerably across industries, from approximately $90,000 per hour in the media sector to about $6.48 million per hour for large online brokerages. Downtime costs also vary significantly within industries. Business size is the most obvious factor, but it is not the only one. For instance, companies that can revert to manual processing can continue to function when their systems are unavailable, although usually at an appreciably lower level of activity. In contrast, some companies, such as online retailers, cannot conduct any business during system downtime. Similarly, certain manufacturing businesses must destroy all of the work in progress (such as food and pharmaceuticals).
To gain a firm understanding of the financial impact of downtime, examine the various subcomponents of downtime costs: revenue, human resources, regulatory and compliance, remedial and reputation impact.
Almost all companies now use IT to record sales. Some organizations, such as low-volume, high-value manufacturers, may still be able to take and fulfill orders manually. However, most retailers and high-volume manufacturers can no longer complete the sales process when their systems are unavailable. This is especially true for Web-only stores whose entire order-taking capability stops when their systems are offline.
A companys average sales per hour, adjusted to account for the percentage of sales that can and likely will be completed manually during a system outage, can be used to estimate the revenue that will be lost per hour of downtime. However, this yields only a very rough approximation.
Not all of the sales that would have otherwise occurred during a downtime event will be permanently lost. Some persistent customers will wait for their preferred vendors systems to come back online rather than switching to a competitor.
If an outage creates a disruption in a supply chain with a high level of expectation in responsiveness (i.e., medical services or overnight delivery), the business may be exposed to damages. Often, damages stem from the inability to deliver (i.e., loss in delivery fees due to arriving late or lawsuits due to collateral damages). These highly publicized situations can impact shareholder value.
On the other hand, the value lost when loyal (or potentially loyal) customers do switch to a competitor is more than just the value of the purchases that would have been made at that time. It amounts to the lifetime purchases those customers might have made.
Human Resource Impact
Another significant downtime cost is lost employee productivity, which can be measured in terms of the salaries, wages and benefits of idled people. If, for example, the work of 1,000 people earning an average of $60 per hour, including benefits, depends on a system that becomes unavailable for one hour, the value of the lost productivity can be estimated as $60,000 for that hour. The estimate might be somewhat lower if some employees are still able to do a portion of their work manually or switch to tasks that require only systems that are still available. Nonetheless, many organizations have become so dependent on IT that an unavailable system totally idles a significant portion of the enterprise, when the old manual processes often no longer exist. For instance, What happens when email goes down? What happens to that new cost-saving voice-over Internet protocol system when the LAN goes down?
Subject to the above caveats, the following is a quick formula for estimating hourly downtime labor costs:
Hourly Labor Cost = P x A x C
P = number of people affected A = average percentage they are affected C = average employee cost (salaries or wages + benefits)
Regulatory and Contract Compliance Impact
Increasingly stringent regulations require organizations to safeguard the availability, reliability and privacy of financial, human resources and customer data. In addition to government regulations, many companies now include service level agreements in the contracts they sign with their customers and business partners. Failure to fulfill the conditions of an SLA can lead to the imposition of substantial penalties. In addition, the regulatory and contractual fines that may result from an inability to perform business functions vary widely, but can be substantial.
After a downtime event, remedial actions are often required to repair the damage. For example, employees might work overtime at overtime rates or temporary staff may be contracted to recover lost data and enter accumulated paper transactions. And, if customer satisfaction was damaged, a costly special marketing program may be necessary to win back customers.
If a company repeatedly misses deadlines or is unable to answer queries from customers because systems are unavailable, customers and prospective customers will lose confidence quickly. Whats more, thanks to todays ubiquitous online forums and social networking venues, the probability that any damage to a companys reputation will spread widely and rapidly is higher than ever before.
Once all of the financial impact subcomponents have been analyzed, adding them up to derive a total hourly downtime cost is simple. Multiplying this value by an estimate of the annual hours of downtime provides a projection of the impact that downtime will have on an organizations profit for the year. In the worst case scenario of a major public relations disaster, the results can be sobering.
Clearly, this calculation is oversimplified. For one, planned and unplanned downtime should be treated separately. While planned downtime can be scheduled for a time when it will least hamper operations, unplanned downtime can happen at any time, including when it will do the most damage.
Forecasts of planned downtime can be based on past experience. Database backups and reorganizations, software and hardware upgrades and other maintenance activities tend to occur fairly regularly. Thus, the average annual amount of planned downtime over the past few years provides a reasonable forecast of planned downtime during the upcoming year.
Projecting unplanned downtime is considerably more problematic. The most costly types of unplanned downtime, such as total data center destruction due to natural disasters or terrorism, are exceptionally rare, but still need to be considered. Other unplanned downtime events, such as a hardware failure or power outage, occur somewhat more frequently. Because the timing of these events is random, data culled from a single organizations experience may not be a reliable predictor of these events future frequency and duration. Consequently, industry averages should be considered and, where appropriate, used to adjust internal statistics on unplanned downtime.
System Downtime and Business Downtime are Not Synonymous
Unlike system downtime, business downtime is not inevitable. Technologies exist that can keep a business functioning when an individual hardware or software component becomes unavailable. For example, high availability software maintains real-time replicas of critical systems and data, possibly at a remote location. When a system becomes unavailable due to maintenance or an unexpected event, users can be switched to the backup system thus eliminating the impact of the failure upon the business. Currently, high availability technologies are mature and robust.
Downtime reduction solutions are not free. Furthermore, the price tends to increase the closer a solution comes to promising near-zero downtime. The only way to make a rational, business case-based decision on the appropriate level of investment in availability solutions is to first understand the financial impact and exposure that downtime has on the organizations bottom line. Carrying this analysis out before an outage happens will yield the maximum benefit to the business and its stakeholders.