Sales forecasting is the technique of where its company’s sales records from few past years. These forecasting models can be utilized to predict the short-term or long-term sales performance of the company in future. Forecasting is one of the very important parts of proper financial planning. Financial planning includes the sales, target, supplies, workforce and so on. A good forecasting of sales can help a company to achieve maximum performance and profit. There are many techniques to forecast sales which are mathematical and non-mathematical techniques. Non-mathematical techniques often are the favourite choice of companies because it is easy to understand and easy to construct.
Time series are known as recordings of processes which change over time. A recording that can be a continuous trace or a set of discrete observations (Ihaka, 2005). There are several types of time series model including AR, MA, ARMA, ARIMA and much more. ARIMA model stands for autoregressive integrated moving average and there are three parameters which are p, d, q. Arima models are well known and flexible class of forecasting model. This model uses past information to make a prediction. This type of ARIMA model is a basic step that can be used as a first principle for the more complex model. In this study, the sales of a restaurant will be forecasted by using ARIMA model.
1.3 Problem Statement
New businesses must make assumptions based on the market rates and good judgment. As a new restaurant, the owner faces few difficulties in managing the production of food, unexpected cash flows and controlling the inventories. The restaurant often faces overstock or out of stock situations. This inventory problem eventually affects the production of food where the supply doesn’t meet the demand. The owner unable to predict the demand of food needs to be prepared which in another word the owner unable to predict the sales.
Thus, through this study, a good ARIMA model can be built to forecast the sales which eventually will help the owner of the restaurant to overcome the problems. ARIMA models considered the technical approach which had been widely applied to forecast financial and nonfinancial time series. Reviewing the existing literature, it is observed that time series model is useful in present-day research for sales forecasting.
1.4 Objectives of Study
The aims of this study are:
a. To analyse the sales data of restaurant using ARIMA model.
b. To build a suitable sales forecasting model by using ARIMA model.
c. To forecast the sales from the ARIMA model
d. To evaluate the forecasting performance of ARIMA model
1.5 Scope of Study
The sales data will be collected from ‘Sarala Unavagam’ (restaurant based on Indian food) in Skudai, Johor. Hundred days sales data will be collected from the mid of December 2017 to the end of March 2018 and will be analysed using ARIMA to obtain sales forecast model.
1.6 Significance of Study
In this study, time series model such as ARIMA model will be used to forecast the sales. An appropriate forecast model is essential to predict future sales figures. Sales forecast is very important part of the financial planning of a business.
Sales data from the past is used to intelligently to predict the future performance. The sales forecast is a good estimate for the demand of food and it will help the owner of the restaurant to manage the inventories where the restaurant can avoid both overstock and out of stock situations. This will provide a better control over the supply chain. It helps the restaurant to prevent losses by making a proper decision based on relevant information. Sales forecasting may be a continuous method wherever it is accustomed improve the accuracy, the owner will improve all aspects of the business performance
Organisations which will produce top quality and correct forecasts are able to see what interventions are needed to fulfil their business performance targets. Besides, this study applies the technical approach to time series modelling instead of the fundamental approach to forecast sales because the fundamental approach is extraordinarily time-consuming. Technical method studies the historical sales activity to predict what will happen in the future. This method of predicting the future is successfully used in several applications include both financial and non-financial. Thus, in this study, ARIMA model will be used as a method of predicting the future values of sales.
Forecasting is the key element of financial and managerial decision making. The common financial time series such as sales forecasting are known to be complex, difficult for econometric modelling, non-stationary, volatile and unpredictable. Regardless this complexity, ARIMA model is expected to give a good sales forecasting model to overcome the problem by achieving the objectives.