There’s sort of this ongoing battle at work about percents. Should you use them or not? For example, is it better to say we had 50 requirements defects this month or that 10% of all our defects this month were requirements?
I’m not sure it’s the absolute right answer, but I now have a philosophy. Percents should never be used to talk about rate of change. For example, if someone says “we had a 100% increase in defects from last month” you don’t know if that means you went from one thousand defects to two thousand or that you went from one defect to two. Both are 100% increases. We often see this in sales literature – “realize a 50% gain in efficiency!” But how much effort were you putting into it in the first place? 50% might be a lot, or it might be a large proportion of nothing.
I like percents for comparing groups of things. What percent of defects were requirements vs. code vs. design? Think pareto charts. I also like breaking things down by percent when looking at groups over time because each month might have a different number of defects, sometimes quite significantly, and using the percentages helps expose whether the groups are stable over time or not.
But not so much for change month to month. I think that just ends up being misleading.
Posted on May 16, 2010
0