5 pts.
 Comparing improvement in software response time with its previous version.
Hi,
I have a question, which looks silly but I am getting different opinion among our stake holders.  I need to report an improved performance in our new version of software, where the total time required for for a functionality, say 'Login' in new version takes 2 seconds where as it takes 4 seconds in our previous version.  
Now, how do I present this in terms of percentage improvement in performance?  Is it ((2-4)/2 * 100 = 100% gain) or (4-2/4 * 100 = 50% gain in terms of time saved) ?
Pls note : This is comparison of same software between two versions and not b/w two different products.  So, is there any rule on which one should go to denominator?.
Any articles you can point me to?


Software/Hardware used:
Manual Testing
ASKED: Feb 14, 2012  6:59 AM GMT
UPDATED: February 27, 2012  8:45:00 PM GMT
5 pts.

Answer Wiki:
Last Wiki Answer Submitted:  Feb 14, 2012  6:59 AM (GMT)  by  Gnarayana   5 pts.
To see other answers submitted to the Answer Wiki View Answer History.
Discuss This Question:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _




 

Is it ((2-4)/2 * 100 = 100% gain) or (4-2/4 * 100 = 50% gain in terms of time saved) ?

Instead using 4 seconds down to 2 seconds, think about 4 seconds down to 1 second or down to 3 seconds. Plug those numbers in instead of 2 and see what the percentages look like.

If you are thinking of an improvement over 4 seconds, then 2 seconds of new response time means that 2 seconds was trimmed off of the original time. If you have a 4 pound roast and you cut 2 pounds off of it, you cut off 50%.

But if you are thinking of something like throughput, then the rate has increased by 100%. That is, instead of doing one ‘Login’ in 4 seconds, now you can do two. The throughput rate has doubled; or it has increased by 100%.

It’s not so much a matter of your formula as it’s a matter of what you’re measuring.

Tom

 66,925 pts.