Econ: Projecting the demand of a business

[GoCubsGo]

Ok. So when the going gets tough and I don't have the answer (most of the time) I have to ask. Reading threads about the colleges and degrees people are getting impresses me, so now I'm going to toss this out there and see if anyone can help.

I would like to see if I cannot estimate the demand of a business and I believe I'd want to use the simple linear regression model. BUT what I do not know is what values I should take for X and Y. Obviously I'd want to use the most relevant data to compare. I thought I would use say the number of new store openings versus the net sales, but the number of new store openings is <=2. I don't know if this is good enough. It's relevant, but maybe not enough. A larger number is with the franchises developed so I decided to take that and I wanted to see the thoughts on the data.

I'm not using real numbers for the company, but I am using numbers that are around what the company is reporting and I do not think it's telling me what I want to know. I'm probably doing that wrong.

Is this not an appropriate way to estimate the demand for a business?

 

[n-abounds]

I'm not quite sure what you're doing here, but I'll give you the best I can...
I don't quite get how this shows demand- at best, it shows supply. Doesn't this just show how sales increase with the #stores, as in with more sales the company is willing to supply more stores? Demand usually relates quantity-demanded and prices... I don't think you want to be saying "at a price of $4000, people demand 4 stores". Is there any one product the company sells that you could chart demand for? If you really want to get complicated you could do this for every one of their products and then aggregate them...

If this is the path you want to go, shouldn't one of the axes be cumulative stores opened (ex. numbered 1-20)...as in with 1 store demand was $2000 (or whatever), and with two stores it was $4200 (or something)?

I guess I'm just thoroughly confused. Did you mean supply? Currently, even if that is what you were going for, isn't that marginal demand because you are using each additional store, not a cumulative quantity?

Possibly you were going for how a new store opening affects sales. This would likely shift the demand curve to the right, but before you do that you need a model for demand as a function of price. It gets complicated, because with each new store, the company captures more of the market (although not every new market has the same demand since they have diff. incomes, tastes, etc.). It would be fairly hard to estimate this impact I assume, because it would shift the curve to the right but also affect the slope. I don't know how technical you want...
Also possibly you just need a model for Sales vs. #Stores (which is more business-oriented, not econ. demand). In this case it'd be helpful to know the total sales and the total #stores up until the time period that you have in your table. Then you can chart how a new store affects sales- and use your stats for a prediction interval (or whatever- it's been a while)

 

[GoCubsGo]

Well yeah, you're right about the supply bit. I should find something, their main product and then go from there. And you're not confused, I'm confused so I don't explain it all too well.

 

[n-abounds]

It's probably hard to find out about their products since you would need to know how much of the product they sold at different quantities- and they for sure haven't tested demand at EVERY price (as in they don't lower their price to $0.01 just to see how much they'd sell). Also, they'd prob. be hesitant to share any of that info with you. So any demand that you'd derive would be an estimate outside the small interval of info that you have- although you said you want to assume linear demand.

But if it's econ demand and not just sales vs. #stores you need to get info on the quantity that they sell at each price.

Then again, I'm not exactly a microecon. oracle...

Demand is notoriously hard to predict...if a company can perfectly predict demand, well...they'd make a lot of money.