Geeks With Blogs

Geekus Con Livus Malcolm Anderson's home for Geeks With Lives

I came across an article last night, discussing the value of agile engineering practices (continuous integration, paired programming, one touch deployment, and test driven design).  In this article, the author discussed how much it costs a company (on average) to fix one bug that has escaped out to production.


The company in question had come up with the number $12,605.


The article did not fully break out where that $12,605 dollar figure was coming from, but I’m assuming it was calculated in one of two ways.


The first, add up the following (adding in additional lines for things that you think I missed).


X number of lost user productivity hours, dealing with the bug, and finally submitting a ticket, followed by trying to track the ticket.

X number of management hours dealing with everything from approving the bugs status to “we will fix it”, to “the users say, the fix has been implanted”.

X number of developer hours understanding the nature of the bug.

1 - 10 minutes to fix the bug.

X number of QA hours confirming that the bug is truly fixed.

X number of operations hours to get the application deployed back out to the users

X number of lost user productivity relearning how use the new correct functionality.


Add in the actual cost of an employee’s time (taxes, insurance, vacation time, training time, real estate costs, etc.)  the employee “fully loaded” cost is usually in the neighborhood of double or triple their salary, but use 2x to simplify the exercise (but I think 3x is more accurate).



The second, simpler way of coming up with the cost of a bug, is to take the IT department’s budget,

divide that dollar number by the total number of bugs fixed for that budget duration

and then multiply that number by the percentage of the organizations time spent on bug fixes, verses new development.


My questions are:

Are you in a position to track your “Cost Per Bug Fixed”?

If you are not in a position to track that cost, how could you put yourselves into that position?


The reason I ask is the following:

Most companies expend little or no thought, on Total Cost of Development (TCD). 

Once the project budget has been spent, this is forever thought of, as that applications “cost of development”, 

I’d like to suggest that this is more accurately calculated by taking the original project costs, and then adding the cost of every bug ever fixed on that application, producing a “Current Cost of Development” (CCD). 


This level of transparency could be used to give product owners more control on their development costs, by allowing them to make a decision that they are already used to making; “Do I want to pay cash, or put it on that 30% credit card?”

In my experience, companies have chosen the usurious credit card because they don't know any other way.

"I'll put this on my 30% credit card" is normally pronounced, "What is the earliest date, you can not prove you can not deliver by"

What's been your experience?

Posted on Wednesday, June 6, 2007 5:42 PM Agile Development | Back to top

Comments on this post: Forget TCO, what it your applications "Current Cost of Development"?

No comments posted yet.
Your comment:
 (will show your gravatar)

Copyright © Malcolm Anderson | Powered by: