Expert: Sherry Wallin - 11/21/2011

Question
QUESTION: I’m looking to figure out an equation that can forecast sales. For example let’s say I have sales for the first five days. Lets also say there are 22 days in the business month, 30 regular days in the month. How can I predict with a certain level of confidence a sales figure?

So let’s use these figures for the example:

Sales by Day
Day 1 25,000
Day 2 30,000
Day 3 28,000
Day 4 27,000
Day 5 30,000

Sales Month to Date
Day 1 25,000
Day 2 55,000
Day 3 83,000
Day 4 110,000
Day 5 140,000

ANSWER: Tom~

Is this a statistics problem? When you say with a certain level of confidence it smacks of statistics. You have a mean of 28,000 which can be found by taking the 5th days sales month to date and dividing by 5: 140,000/5 = 28,000. You could then calculate the margin of error and a get a confidence interval. Since you don't have a population standard deviation you would have t calculate the sample standard deviation using sqrt(n*Sum(x^2)- [Sum(x)]^2) and then using the students t distribution with 4 degrees of freedom at whatever confidence level you choose using E = t_(alpha/2)*s/sqrt(n) and then look at the confidence interval
x bar - E < mu < x bar + E

Math Prof

---------- FOLLOW-UP ----------

QUESTION: I think I may have missed some important points in this quesiton but first thanks for the work already done. So let me ask this question again with different figures and in a different way.

My goal here is to predict sales at the end of the month. So below is a breakdown of sales for 10 months by week. How can I create an equation that at any given time can predict what sales will be at the end of the month. In otherwords, lets say it is October 10 and I want to estimate what the monthly sales will be as of October 31. I attached a file with the information.

Answer
Tom~
There are a lot of things that will affect sales but in general if you have data for the previous years in the month of October you could use that average to predict what you will have in sales this year in October. You could even take the average of what is already made in October, i.e., add up the sales you already have in October and divide by the number of sales days and multiply that by the number of sales days left in October and add that figure to whatever sales figure you already have for October. This should bring you faily close to what your sales will be in October.

Math Prof