IT organizations across the globe use a number of metrics to measure their successes, failures, and standings. One of the more popular metrics is the “number of 9s” as a measure of system uptime. Why use 9s? It is relatively easy for technology organizations to measure system performance. Unfortunately, it does not matter outside of IT.
What are 9s?
The number of 9s refers to the percentage of system uptime. Typically, we hear about three 9s, four 9s or five 9s. Three 9’s refers to 99.9 percent uptime, or .1 percent downtime whereas five 9s refers to an ever-illusive 99.999 percent uptime or a mere .001 percent downtime.
These metrics have been used for a very long time; from internal IT organizations reporting status to Service Level Agreements (SLAs) from service providers. The number of 9s is used as a metric to set performance target . . . and measure progress toward them. The problem is, they are technology focused. When looking at the inverse as a function of downtime, it equates to the following table:
Even at four-9s, that equates to a maximum of only 52.56 minutes of downtime per year. Unfortunately, this means very little if the company is in retail and those 52 minutes of downtime came during Black Friday or Cyber Monday. In addition, the number may be artificially low as other factors may not be included in the calculation.
The fallacy of planned vs. unplanned downtime
First, it is important to differentiate between scheduled downtime and unplanned downtime (outages). Most measure their system performance based on the amount of unplanned downtime and exclude any scheduled downtime from the calculations. There has been an ongoing debate for years whether to include scheduled downtime.
Arguably, if a system is down (planned or unplanned), it is still down and unavailable. In today’s world of 24×7, 100 percent uptime expectations, planned downtime must be considered. Ironically, the inclusion of planned downtime causes uptime figures to drop and may cause a rethinking of how applications and services are architected.
In today’s world, do these metrics even make sense anymore? They are not business metric . . . unless you are a service provider that makes your business about uptime. For the majority of IT organizations, these metrics are just “technology” metrics that have little to no relevance to the business at hand. Just ask a line of business owner what five-9s means to their line of business. For IT, it is hard to connect the dots between percentage uptime and true business impact. And by business impact, this refers to business impact measured in dollars.
If not 9s, what business metrics should IT be focused on? Most companies use a common set of metrics to gauge business progress. Those may include Cost to Acquire a Customer (CAC), Lifetime Value of Customer (LVC) and Gross Margin. Customer engagement is a key area of focus that includes customer acquisition, retention & churn. For IT, these metrics may seem very foreign. However, to a company, they are very real. Increasingly so, IT must connect the dots between that new technology and the value it brings to business metrics. As IT evolves to a business focused organization, so should their metrics of success.
The role of the CIO
The CIO, above all others, is best positioned to take the lead in this transformation. Instead of looking for ways to express technological impact, look for ways to express business impact. It may seem like a subtle change in nomenclature, but the impact is huge. Business metrics provide a single view that all parts of a company can directly work toward improving.
A good starting point is to understand how the company makes money. Start with reading the income statement, balance sheet and cash flow statement. Are there any hotspots that IT can contribute to? And what (business) metrics should IT use to measure their progress.
Not only will this shift IT thinking to be business focused, it will also highlight better alignment with other business leaders across the company.