Sales Research and data extrapolation with sales questionaire - HELP !

[ram_sankar]

Question for Sales:

I am working on a troubling research to find what attributes of a company contributes to sales. This is related to my master thesis for MBA. Thus a sales questionnaire we circulated helped us understand what are the top 12 reasons why a customer buys a product. The sales staff were requested to rank the questions based on their experience, what are the top important reasons why a customers buys a product(1 - best reason and so on till 12). Table below details the questionnaire outcome:

Q.No	Question	Sales Staff 1	Sales Staff 2	Sales Staff 3	Sale Staff 4	Sales Staff 5
1	Customers bought product X due to Product Quality(Made in Germany - Trust)	3	2	4	6	1
2	Customers bought product X due to better discounts offered than competitors	10	5	13	13	6
3	Customers bought product X due to Aftersales Services	1	3	2	1	2
4	Customers bought product X due to Product features	1	1	3	1	1
5	Customers bought product X due to Existing good relationship	1	2	3	1	1
6	Customers bought product X due to Contractual Obligations such as frameagreement etc. by central purchaser	9	10	12	10	6
7	Customers bought product X due to International Presence of Product X manufacturer	9	8	3	7	6
8	Customers bought product X due to international counterparts used the same brand in other country operations	9	8	5	10	2
9	Customers bought product X due to previous good experience	2	3	2	3	1
10	Customers bought product X due to personal relationship between Buyer and product X manufacturer	1	2	3	1	2
11	Customers bought product X due to competitors product bad experience	8	10	8	8	7
12	Customers bought product X due to customizable solutions offered by manufacturer than other competitors	3	4	1	1	6
Total No of Sales for year 2014 :	100000 units

Now my task is to calculate how to arrive at, what % of sales is attributed to question 1, 2, 3,…12 to be able to come to a conclusion which are the main attributes constitutes to sales.

I know one way is to calculate the average of all the questions and then sum them to take an individual % of each(but this has to be done with the reverse ranking where 12 will be the best score). But is this the accurate way or is there some other models that would be better suited.

You quick response will be immensely appreciated.

Thanks a million in advance for your experienced suggestions. :cool:

 

[stapel]

What does this have to do with "new information or announcements regarding the forums"?

 

[Ishuda]

...I know one way is to calculate the average of all the questions and then sum them to take an individual % of each(but this has to be done with the reverse ranking where 12 will be the best score). But is this the accurate way or is there some other models that would be better suited.

You quick response will be immensely appreciated.

Thanks a million in advance for your experienced suggestions. :cool:

There are many ways to form an 'average'. The arithmetic average, harmonic average, a weighted arithmetic, etc. In your case I can't see a reason to do other that an arithmetic weighted average:
Let w_i be a set of weights about how much we think the i^th Sales Staff response should be counted/believed/weighted and W be the sum of those weights. Then you would form the weighted average for each question, i.e. for question 1 we would have a score of
S₁ = [3 w₁ + 2 w₂ + 4 w₃ + 6 w₄ + 1 w₅ ] / W
Given no reason to weight one Sales Staff response over another, you let each w_i=1 and have the usual average. Given sales figures or some other measure for each of the staff maybe a weight based (at least partially) on that.

 

[stapel]

Moved from "News" to "Finance / Business Math".