0

I want to compare various scores with one another on a fair basis. Let's say I have 4 results:

  1. 10/20 (50%)
  2. 1/2 (50%)
  3. 1/4 (25%)
  4. 2/10 (20%)

I need to assign weights of importance to the parameters individually before rebasing it to a unified score out of 100. For example the first two scores result in the same % value. But the first score should have a better final score, as the demoniator indicates better performance in my usecase.

Mathematics is not my speciality and I was wondering what the best way to do this would be?

UPDATE

The parameters above are: (Sales Realized)/(Sales Promised)

In order to make it more fair, I should probably account for the number of calls as a parameter as well. Am I right?

HermannHH
  • 101
  • Interesting. Why do you say Score 1 is better than score 2? Can you quantify your preference at all? (e.g. is 9/20 better or worse than 1/2?)? – lulu Jul 10 '15 at 12:21
  • @lulu this is for a sales call problem. In the first case the agent made way more calls. Even though his % is lower when a normal % is calculated. In your example, I would say that the first case should be better. You made me think now... I have updated my question accordingly. :) – HermannHH Jul 10 '15 at 12:24
  • What is "sales promised"? Who "promises" a sale, and why do we care about this and not just about the number of sales that are realized in the end? – David K Jul 10 '15 at 13:08
  • @DavidK, Sales promised is a indicator of how effective the agent was. One could argue that closing the sale should be the sole metric used but, after the sale is "promised" the customer is handed over to another department. – HermannHH Jul 10 '15 at 13:12
  • Ultimately, doesn't the success of the business depend on how many good "promised sales" the agents hand over to the other department--namely, the number of those sales that are realized? If anything, the promises sales that aren't realized are costing the company resources (labor and materials expended on them by the other department) and might count against the agent; but it depends on how much they really cost, compared to the gain when a sale is realized, so it might take many unrealized promises to cancel the benefit of one realized sale. – David K Jul 10 '15 at 13:22
  • Are you asking for a method to generate some type of performance indicator based on the ratio of (Sales Realized)/(Sales Promised), with an increase in this value as the totals of both, Sales Realized and Sales Promised increases. – Peter Jul 10 '15 at 14:18
  • @Peter t that is exactly what I'm trying to arrive at – HermannHH Jul 10 '15 at 14:19

0 Answers0