I need to create a scoring system for some Widgets. I am looking for
the top and bottom performers in a population of Widgets. The data is
in excel, and this is what I have:
Widget Name, Scale Score (1 is lowest and 5 is the max), Frequency of responce
Widget A, 2, 1
Widget A, 5, 1
Widget A, 1, 1
...
Widget B, 5, 1
Widget B, 2, 1
...
Widget C, 5, 1
etc...
I am able to subtotal the sheet with sum, by Widget. This gives me,
per Widget, the total Scale score and how many ratings for that
Widget. I then take the average for each widget type. But then I am
stuck. I need a scoring system that takes into account the average
score AND the number of answers for the Widget type. I tried doing a
weighted average where the weight is the % of frequency, but that gets
me no where. The distribution of the scale score is normal, but the
frequency is highly skewed where a bunch of the widgets, say Widget A
and Widget B, get most of responces.
So if Widget B gets 30% of responces and its average score is 3.2 it
shouldn't necessarely end up in the top 10 widget, if other widgets
with better average score but only 1-2% of responces excist.
I would like a mathematical formula that takes the averages and count
frequencies, and tell me what the top 10 performers are and what the
bottom 10 are.
The data in the sheet is done weekly, and I would like to report on it
weekly with a Scoring System that acts like an Index (S&P 500, Nasdaq,
etc (it doesn't have to be, that's just the way I think of what it
could be like)) for each widget. New widgets are introduced all the
time and I want to see rightaway if they end up in the bottom or top
10 after just a few weeks.
Any ideas?
Thank you |