# Comparison of Multiple Groups of Data...

#### bubbisthedog

##### New Member
Man, what a killer forum this is. Thanks to those who are providing it... I'm one of those guys who needs brushing up on the fundamentals, and then it should be relatively smooth sailing.

I have 12 marketing reps' net earnings listed for customers that they represent. So the number of customers per marketing rep is variable. I'm trying to see how one rep's figures compare with the rest of the group. For example, consider only 3 reps:

Sarah | Bob | Jim
110 | 100 | 30
109 | 90 | 25
106 | 80 | 23
100 | 70 | 12
87 | 60 | 11
75 | 50 | 10
| 40 | 9
| 30 | 9
| 20 |
| 10 |

How do Jim's figures compare with Bob's and Sarah's? Thank you for insight that you can provide.

Regards,

bubbis

#### JohnM

##### TS Contributor
As is usual in the consulting forum, I'll answer your questions, at least initially, with another question:

-What aspects of each rep's figures do you consider important? the sum? the average? the number of reps? the variation?

So, it actually depends on what is important to you - then that will lead to the appropriate comparisons.

#### bubbisthedog

##### New Member
Thanks for the response, John!

If averages provide good representation, then that would be great. Of course, I'm not exactly sure how you're going to employ it quite yet. hahaha

Regards,

bubbis

#### JohnM

##### TS Contributor
How "I'm" going to employ it?

Just using the eyeball method tells me that Jim's numbers are a lot lower than Sarah's or Bob's. However, I don't know anything about the particular situation (i.e., is there a rational explanation for the difference, such as difficult or small accounts to begin with, or is Jim just a slacker?).

Averages are only as representative as the underlying data allows....if it's highly variable or skewed, then the average could be very misleading.....

#### bubbisthedog

##### New Member
How "I'm" going to employ it?
Thanks again for the insight, John. I hope you didn't think that I was expecting you to do something! hahaha What I meant was that I didn't know for sure whether to use average, variance, etc. because I wasn't sure what your approach would be depending on which one I chose.

I'll attempt to explain clearer: My reps need to focus on the top 20% of their customers, based on net earnings (applying the Pareto Principle, which you were so kind to refer me to the other day). So, if a rep represents 100 customers, then 20 of them will be the customers that they need to focus on. But if one rep's (Jim's, let's say) top earners are too low to begin with, then Jim doesn't need to be focusing on his top earners, but rather focus on attaining different customers. With the example that I provided, it's easy to see; but with my real sets of data, each rep has hundreds of customers. So I was wondering if there was an analytical tool that tells me where each rep stands in relation to the other reps (according to their net earnings).

I hope that's more clear, and thank you, John, for being patient with me.

Regards,

bubbis

#### JohnM

##### TS Contributor
I would start with something exploratory, like box-and-whisker plots - this will give you a sense for the average, dispersion, and any "outliers" for each rep.

Also- look at the number of customers for each rep so that you view each rep's plot in the proper perspective.

I think the patterns you see in these plots will lead you toward what to do next....

Unfortunately, there's really no "magic" method that will produce what you're looking for on the first shot - it takes some digging before you find the gold nugget of insight.

#### bubbisthedog

##### New Member
Yeah, I couldn't recall a way to do it, so I just wanted to make sure I wasn't forgetting anything. I'll definitely use the b-a-w plots and do some digging to make my decisions.

Thank you, John, for taking the time to help me and providing valuable insight! Now that I've found this excellent forum, I'm sure you'll be seeing plenty of posts from me!

Regards,

bubbis