As most of you are probably not aware, most of my career has been spent in security, Disaster Recover, High Availability and infrastructure/datacenter planning. A big part of this in designed systems is what type of performance and availability do you need to plan for. The more available a system needs to be the more expensive it will be to get on line by using such things as clustering, duplicate storage, etc. This is usually because downtime costs money. Calculating how much you should spend on making a system highly available is pretty easy, calculating ROI is a breeze is a matter of figuring out the answer to the question…How much does downtime cost?

For example I worked for a company that had a significant investment in call centers. The call center operated 24x7x365 and the phone was ringing pretty much constantly. During the consulting process I worked on, the cost of downtime was estimated at $245,000 per minute. So if we wanted to spend $1million dollars that would guarantee reduction of downtime by 20 minutes per year. The ROI was simple.

Unfortunately ROI and NPV are not always based on reducing downtime. Usually with regards to applications the question is how can we justify buying this or that? I think those of us in technology hear this a great deal and have often struggled to really get across the value to our management in order to get expenditure or projects approved.

This weekend it occurred to me, is making a process more automated and more efficient any different than stopping a system from being down. If you either make a system faster by 1 minute or you prevent it being down for 1 minute the end result is basically the same.

So how do you estimate the value of making a process more efficient? Here is what I thought of.

Let's take the fully loaded average cost of an employee. For example sake let's just say that its $50,000

And that worker works an average of 46 weeks in a year. Or 230 work days.

This totals about $1086 per week

• $217 per day
• $27 per hour
• .45c per minute.

When taking this by itself it does not seem like much, and you are probably thinking how does this help me? Let's take this a step further. If you automate a process that helps you be more efficient, and you gain insight into your processes that continually make it more so. Even 1 minute improvement makes a difference. How you ask?

Well 1 minute of that employee we talked about earlier is worth $.45. If you improve a process that saves that employee 1 minute per day and there are 340 work days in a year you made a savings of $103.50 per employee that was involved in that process.

So take one of your processes that are rather wide spread, then figure out how many people are involved with it. Take an honest look at the process and try to gauge what some realistic optimizations would look like if you properly automated it and implemented some kind of BPM with it. Could you save a few minutes per process instance, 1 minute per process instance etc?

Now calculate out like I did the fully loaded costs and break it down just like I did. This will allow you to pretty easily figure out what kind of cost savings you can deliver. I am experimenting with a spreadsheet that shows this example and tries to do a little more, when I finish I will post it.

I do realize that there are many other factors that can be used as additional justifications for purchasing software, implementing new systems etc. Such this as happier workers, can you put a value on users that are generally happier? Can you put a price on users that are more able to help themselves? It has been pretty difficult to accurately estimate this type of benefit. But maybe someone has good ways to do that. Maybe you can even write a process to adequately survey each team member J