Business Analysis and Decision Making

**Introduction**

Many firms always have difficult time of decision on some of the tools they can employ to come up with better decisions that can improve their sales. Most a times, management of a firm shall always hold meetings in with stake holders shall give various views why one option is better than the other. Business managers seldom think of any computer aided decision making tools which can help a company to come up with better decisions that can help them to predetermine the status of the sales in future based on any independent parameters that the sales depend on. Even though there are a number of options, linear regression stands out be the best statistical option that a business can put in practice to predict the quantity of sales its likely to make in future based on array of factors. This write-up shall use the number of hits on a webpage to predict the value of sales the business is likely to have in the next three months.

Therefore, the write-up shall help to come up with records that shall help the client to determine the quantity of sales the client shall make daily in the future. This forecast shall be obtained using statistical linear regression method that shall be drawn using Excel spreadsheet.

**Linear Regression Data Analysis**

Regression curves are often used as an approach of visually illustrating the relationship between the independent (x) and dependent (y) variables so that one can easily conclude whether the two variables are related or not (Wilson, Keating & Galt, 2012). In addition, this research also uses the co-efficient of correlation, also called the R-square to determine whether the forecasted sales values can be relied on or not.

The graph represents a scattered graph of the data provided on sales with equivalent hits that were made. Since Hits are independent of any other variable in this research, they are therefore put on the x-axis. In this research, the sales are dependent on the hits made on the web page. Since they are dependent on the number hits, they are put on the y axis. When a trend line is drawn it gives an equation y = 1.2485x + 519.44 and an R² of 0.5973.

Based on the equation of the trend line, more forecasted values can be obtained as long as the number of hits is obtained. For instance, the following table summarises the outcome of sales in the next three months

x | y | |

Month | Hits | Sales |

Dec | 1287 | 441 |

Jan | 1164 | 1972.694 |

Feb | 1159 | 1966.452 |

Mar | 1298 | 2139.993 |

April —- —- |

The linear Regression equation, y=mx+c, the coefficient m is the multiplication factor that is used with the number of hits to determine the sales. In other words, it defines the number of hits one should have to obtain the value of given sales. C is a constant or y-intercept which defines the minimum sales the seller even if no hit has been recorded on the website. Therefore, the equation is a true reflection of what value of sales are likely to be obtained, even without using extrapolation method on the graph which utilizes the position of sales on the line based on a given number of hits.

For January, 1.2485*1164+ 519.44=1972.694

For February, 1.2485*1159 + 519.44=1966.452

For March, 1.2485*1298+ 519.44=2139.993

**Discussion**

Linear regression line thus is most suitable to determine the future sales based on the current p outcomes. The R squared (R^{2}) is the correlation co-efficient value that determines how real is the data obtained compared on the actual facts. When the R^{2 }is 1, it means the data predicted using the linear regression method is 100% true and the relationship between the hits and the sales is directly proportional. In other words, any value forecasted about the sales based on the hit shall be the actual sales when the time comes. On the other hand, if the value is zero or trending towards zero, then it shows there is no relationship between Sales and hits made on the website. In such a case, it is quite difficult to determine any value of sales in the future.

In this case, the correlation value of R-squared is 0.5973. This shows that the accuracy of the data is 59.73% true. Therefore, one cannot entirely rely on the forecasted values as the true sales that shall be determined in the future based on the number of hits.

Based on this analysis, the client should understand that the findings are true and can be relied on to determine the sales of future months based on the number of hits. However, the data cannot be fully relied on since its R-squared value is not 1 but slightly above 0.5 (Chase, 2013). In other word, the client can adopt this method to determine his future sales since it provides better sales precision compared to any other statistical methods.

Some of the drawbacks with this method are as follows;

Firstly, the method assumes that the prevailing factors shall always be same regardless of change in both internal and external factors that can affect the business. In this case, if the website was hacked and started showing wrong number of clicks, the linear regression method would still take in the values and use them to determine the value of sales, which would be a wrong figure.

**Conclusion**

With aid of excel scattered graphs, one is able to come up with a linear regression graph that can help to predict the sales values. When a trend line is introduced in the graph, future sales can be easily determined based on the current number of hits. The Linear regression also comes with coefficient of the correlation value that determines how true the obtained data is. Therefore, this research has used regression graph to determine sales the client is likely to get January, February and March. However, the method resents a number of drawbacks which have been discussed in detail in the discussion section.

**References**

Chase, C. (2013). Demand-driven forecasting: A structured approach to forecasting. John Wiley & Sons.

National Science Foundation . (2005). Linear Regression in Excel. Labwrite, 1.

Wilson,J, Keating,B & Galt,J. (2012). Business Forecasting. Michigan: McGraw-Hill.