I’ve noticed lately that there seems to be a lot of searching for information on the “cost of downtime”, so I thought I’d spend a few minutes going over some calculations in detail. This item is an expansion of our previous article, which references a Gartner Group study on the cost of downtime.
The math itself is easy, but the tricky part is in identifying the variables to apply. After the jump you’ll find a table you can use to do the calculations yourself. And hopefully convince yourself that having computers go down just isn’t worth it any more.
The labour cost isn’t hard to sort out, if you have the right info. The big variable is the second line, the productivity lost. If the computers are down, do your people sit on their hands? Or can they do at least a part of their job? How much less efficient are they at what they do? 40%? 80%? This is what needs quantifying. And once you do, you know the total labour cost to production.
| Productivity per person in dollars | |
| % reduction in productivity per user | |
| Productivity X %reduction in decimal (e.g. .35) | Cost of lost productivity in $ |
| Avg. Labour cost/hour | |
| Labour cost X %reduction in decimal (e.g. .35) | Cost Per staffing in $ |
| Cost of Productivity + Cost Per staffing | = $ dollar amount |
| Multiply by # staff affected | Cost per department or company |
| Repair Cost for parts/software | |
| ADD Repair Cost Labour (per hour) | Total Repair Cost |
The next big variable set is the loss of good will and/ or repeat orders. Your typical MBA or CMA will probably have these numbers. Me, I’m a geek, and I don’t have them. If a customer phones (for example, and their order can’ be taken, will htey ever come back to order again? or if their order can be taken on paper, but it’s much slower, how do you quantify that negative perception? How much is a bad experience worth in terms of lost future sales, in real dollars?
So let’s do a sample calculation here, shall we? Let’s say that we’re the IT folks for a medium plubming supply wholesaler. The customer service department consists of 10 CSRs (for simplicity’s sake, we’ll leave their manager out of this discussion) who, for reasons unknown cannot connect to their order entry database. We’ll say they they each do $500/ hour in sales, and they each get paid $15.00/hour. While their computerized order entry is down, they’re stuck doing paper order taking, which is about half as fast as computerized order taking.
There is no on-site IT, so management call a tech support company, who take 30 minutes to arrive, another 45 to diagnose, and a further 45 to actually perform the repair (for our purposes, we’ll say that the power supply on a network switch failed, so the switch needs to be replaced. The “repair’ in this instance is to go out to the local shop and acquire a new switch). Actual downtime is 2 hours.
With all that in place, let’s fill in the blanks on the chart above:
| Productivity per person in dollars | $300 |
| % reduction in productivity per user | 50% |
| Productivity X %reduction in decimal (e.g. .35) | Cost of lost productivity in $ 150 |
| Avg. Labour cost/hour | $15.00 |
| Labour cost X %reduction in decimal (e.g. .35) | Cost Per staffing in $ $7.50 |
| Cost of Productivity + Cost Per staffing | = $ dollar amount $157.50 |
| Multiply by # staff affected | Cost per department or company $1575.00 |
| Repair Cost for parts/software | $100 |
| ADD Repair Cost Labour (per hour) | $200 |
| Total Repair Cost | $1875 |
The to this number we add the wildcards I descibed above: goodwill, repeat business, lost sales (if the CSRs spend more time on the phone, there’s a better chance someone calling can get a busy signal, or hang up after being on hold too long).
The result is… wow. Who would have thought that a simple outage like this could end up costing this much money to the business? It’s easy to see why Gartner pegs a large-size company outage at over $42,000 per hour; we’re not even including ALL the variables here (how much productivity is lost in shipping because the orders aren’t as clear? Because they’re slower to arrive? How much good will lost because of order errors as a result?)
The bottom line is that in business, you simply can’t afford unscheduled downtime. WEith two outages like the one described above every year, you can probably be sure to have customers notice, which increases the hit to good will.
And how do you protect yourself? Managed services is one way. Service level agreements are another way. No matter what though, business and IT leaders should do the math, and find a way to keep their technology in check.