Performance Analysis and Forecasting for TourneSol Company

Introduction
When it comes to product production and business, it is always prudent to monitor every step of
production as well as be able to:
i. Minimize the cost of production
ii. Forecast the market behavior
iii. And, trace the performance of the company
Purpose of report
The purpose of this report is to analyze the performance of TourneSol Company using historical
data so as to be able to forecast factors such as average purchase prices of the production inputs
in the coming year i.e. marketing year 16. In addition be able to offer recommendation as to
which supplier to purchase from which % of input so as to minimize the feedstock cost. The
paper uses the Anderson et al (2015) descriptive, predictive and analytics text for application in
excel solver where applicable to enable data-driven decision analysis and making.
Description of problem
In the production of Oil by the company, the maximum amount of iodine required is 0.88% and
the minimum is 0.78% while the minimum Oleic acid that should be in the product should be
77%. Each supplier has a different kind of raw material that has different concentrations of the
required contents. The task is to determine at which percentage of raw material the company
should purchase from each supplier to ensure minimum cost of feedstock while attaining the
required content for quality oil.

Assumptions
In the preparation of cost-volume-profit data, entries such as taxes are obtained from government
source so as to enable an almost real-life business situation (they are subject to change).
Methodology
Forecasting
The method used for forecasting in this report is the exponential smoothening. According to
Otext (2017), “...all future forecasts are equal to the last observed value.”
Hence:
yT+h|T=yT
Where h= 1, 2, 3... which is the naïve method that supposes that only the latest observation is the
most important. In exponential forecasting each forecast is solved through applying weighted
averages such that they decrease in an exponential manner as the historical data gets older
(Makridakis and Hibon, 2013).
yT+1|T=αyT+α(1−α)yT−1+α(1−α)2yT−2+⋯
According to Casey (2018) the triple exponential smoothening takes the form:
It = Β xt/SSt + (1-Β)It-L+m
Where:
 x = observation,
 SS = smoothed observation,

 B = trend factor
 I = seasonal index,
 F = forecast m periods ahead,
 t = time period.
Forecasting data
The forecasting data is extracted from the company’s historical data for the past 15 marketing
years with the 15th year being the current year.
Linear Programming
In linear programming, we have 2 variables i.e. x and y with 3 constraints such that:
Where X denotes amount of Iodine present in each supplier’s raw material and Y is the amount
of Oleic acid available.
So as to obtain the ratios that the company has to purchase from each supplier to minimize the
feedstock cost.
From the initial problem, the least amount of iodine that should be available in the product is
0.78% while that of Oleic acid is 77%. Supplier C has 0.72% iodine and 65% Oleic acid,
contents that are less than the least required by the company’s product.