To estimate new sales from history

A fruit supplier sells 3 types of fruits. The company has 3 salespersons.

Here are the sales quantity of each person for each fruit. The total sales figure is available. (this is all the available data)

Because of different reasons, e.g. season, event, bulk purchase, etc. The unit sales prices were all different (and has no record in place).

In the new year, management wants to estimate how many new (in addition to the old) business these 3 salespersons can bring in.

Assuming each new customer is \$100, How many new customers that David, Mike, and Lilly can generate in the new year?

An idea:

I. to calculate the weight of the sales of each fruit type in 2020
II. multiple the fruit quantity each by the calculated weight (I)
III. Sum all the II of each person
IV. to calculate each person's III's weight in all IIIs.


The IV of each person is the expectations of new customers, i.e. David is expected to bring in 3 customers, Mike will bring 5 etc.

Does this idea make sense? What would be a better choice and why?

• If this is an actual real-life problem you are facing, A+ for obscuring it to make it look like a stereotypical textbook/school assignment problem. "David, Mike and Lilly" - marvelous, simply marvelous. – Mark L. Stone Feb 20 at 13:16
• @MarkL.Stone, thank you for your comment. Yes, it's a real life problem. Just there are more salespersons and more fruit types. – Mark K Feb 20 at 13:26
• @Mark K, as you have faced with the problem that you can estimate each Pearson sales by an estimating function (e.g. moving average, etc), it is interesting to use this function directly or by Monte Carlo simulation if there is some uncertainty in the problem. – A.Omidi Feb 20 at 15:11