Turning sales forecasting, which tends to be a weak point, into a specialty

In store development, sales forecasting for new stores is an area where many people are troubled, saying, "It's hard to get it right" and "I don't know how to do it."

However, this sales forecast is something that cannot be avoided in the presentation of the property, and whether it is removed above or belowserious problemwake up.If you can't do it properly, it's not good for you as a store development manager.

So, in order to be able to forecast sales, I look for books on statistics and Excel and try to do it, but I can't find anything that is directly linked to my work. .

If you don't have one, let's try to make one, so in this blog, I would like to explain the essential knowledge used for sales forecasting in store development in an easy-to-understand manner.

I will talk a little bit about mathematics, but only what I learned by junior high school will come out, so please don't be shy.

This time, as the first time, I would like to try the touching part of the explanation about "regression analysis".

What do you mean by "predict sales"?

Predicting sales in the first place means "measurement of sales of new properties numerically" based on the trends of similar stores such as existing stores and, if there are no such stores, competing stores.

As for how to measure, we analyze the trends of existing stores and similar stores, apply the analysis results to new properties, and produce results.This kind of work is called

Therefore, it is not a so-called "prediction" that if you open it here, it will sell or not.The ``measure'' of prediction is the ``measure'' of measurement.Measuring in numbers is predicting.

Therefore, this requires an objective explanation.An objective explanation means that no matter who does the same thing, the same result will be obtained.

The method of generating this prediction value should not be done in a self-taught manner, but should be done according to proper manners and methods, so it is required to use statistics and data analysis.

Persuasiveness will increase by properly applying statistics and data analysis methods.As a result, it will be easier to obtain the consent of the company, so it is important to acquire such statistical methods and make sales forecasts.

First step: creating a scatterplot

The first thing I would like you to do is create a scatterplot.I will explain it in this figure.

To create a scatterplot, first take the vertical and horizontal axes.In the case of sales forecast for store development, the vertical axis is store sales.The "?" on the horizontal axis is the matter that affects sales, and is also called a "factor" related to sales.

As an important point, it is decided what to put on the horizontal and vertical axes.Please remember that the horizontal axis is the cause, or the "determining factor".The vertical axis is the result.So the vertical axis is store sales.Simply remember that the vertical axis is store sales and the horizontal axis is the factor.

A scatter chart is a plot of existing store data that collects store sales and factors related to sales for each store.

For example, it will look like the figure.There are XNUMX green dots.There are XNUMX existing stores, and the store sales range from high to low. is also going up.You can read this kind of relationship.

In order to visually understand these things, we will create a scatterplot.When doing this in practice, we use Excel.

Then, as the number of "?" increases, the sales of the store will also increase, so if you draw a line that passes between these XNUMX green dots, you will see a red line as shown in the figure. .

What happens if you create a scatterplot and show the two relationships as a straight line?Looking at it is the first step.

Step XNUMX: Do a Regression Analysis

Then comes regression analysis.Regression analysis is a technique that appears in statistics, but I would like to explain it as simply as possible, so please bear with me.

Look again at the scatterplot above.After drawing a scatterplot, I talked about thinking about a straight line that passes exactly between the seven points. Regression analysis is the task of determining that red line as a single line.

First, as a trial, please draw a red line for each of you while looking only at the green dots in the scatter diagram above.Then, I think that the red line will change depending on the person.Even if you are the same person, if you do the same thing again after a while, I don't think it will be exactly the same line.In this way, the way the line is drawn varies from person to person.Therefore, I want to decide on one red line so that it will be the same no matter who sees it or who decides it.

The value of "?" is determined by examining the target property for matters that affect sales.Going up from that value hits the red line.This red line is used to calculate the sales forecast value.Regression analysis is trying to create such a mechanism.

If you decide one value of "?", one predicted sales will be decided.These things are called "functions". When one value of x is determined, one value of y is also determined.As I learned in junior high school mathematics class, such relationships are called functions.It's called a linear function, but it's used in this kind of place even in the work of store development.

Determining one red line means determining one formula: [store sales = certain coefficient x value of "?" + constant].Determining an equation means determining coefficients and constants.This leaves only one red line.Determining this is called regression analysis.Recognize that what you do in regression analysis is to find these coefficients and constants.

I think there was a formula y=ax+b in the linear function you learned in junior high school.To decide this a and b
It is called regression analysis.You may have learned the constants shown here as "intercept" and the coefficients as "slope" and "rate of change".

By determining the function through regression analysis, one line is determined, and one store sales value is determined from the "?" value of the property.If one forecasted sales is determined through the function of the red line, no matter who calculates it, the same number will be obtained.

In this article, I hope you will understand what regression analysis is.

The actual analysis is done in Excel.There is a data analysis function, which is very useful.

Simple regression analysis and multiple regression analysis

Now, as for the "?" which is the factor that affects sales, there are usually more than one. One or two is not enough.

Therefore, there are two types of regression analysis: simple regression analysis and multiple regression analysis.

A simple regression analysis is when there is only one factor.Such cases are rare. Two or more are often related.In this case, we use multiple regression analysis.

The only difference between simple regression analysis and multiple regression analysis is whether there are one or two or more factors, and the basic idea is the same.

For next time: In practice, there are various tips and ways of thinking

In the previous example, there were seven points, and one red line could be drawn neatly, but in reality, it is almost impossible to draw one line neatly.I said that there are usually multiple factors, but it is not so easy to make multiple factors related to everything.

However, there are almost no books that contain only knowledge that can be used immediately at work, so in this blog, I would like to explain how to do it and tips.If you are interested, please contact us again.

So this time, what kind of regression analysis is used for sales forecasting?I tried to talk to you.First of all, it is fine if you can roughly grasp the image.Also, I think it would be good if you could understand that the mathematics you learned in junior high school is actually used in such places.

More on that in the next blog.Our YouTube channelAlsomini lecturehas been uploaded, so please take a look.