Forecasting is not only about applying time series models and performing predictions but also means to assess the performance and make better business decisions. In this blog, instead of focusing on the math behind forecasting, we will focus on how forecasting is interpreted in the Business analytics domain and the different approaches pertaining to it. To start with the definition, business forecasting is a way to predict or estimate the future based on past and present data. Now, this data can be

• Historical data, for example, the stock price of Uber or sales data of Walmart.
• Opinions from experts who have expertise in the domain which might help assess future events and identify the trends.
• Known variables, for example, Amazon keeps track of promotional events based on region to uplift the sales and come up with a better pricing strategy.

---

Why Forecast?

In the Retail Analytics domain, it will help to identify a month or week where we will need more items in inventory. For example, during the new year, most of the retail outlets have more items in stock as compared to the rest of the year.

In the HR Analytics domain, forecasting will be used to identify the period to hire more employees, so that projects get utilized.

In the Production industry, it would help us know as to when to launch a new product or service, expected costs and profits surrounding it so that it would help stakeholders to devise a strategical planning upfront.

---

Forecasting Methods

There are two general methods which Business Analysts use for a forecast which are:

• Quantitative Methods
• Qualitative Methods

Basic Forecasting Approaches

Quantitative methods are associated with historical data(several years of data) and facts associated with it. For example, the stock price of equities in NSE. These methods look at the trend, which will help to answer whether the stock is moving upward or downward. Changes due to population, technology culture, weather are taken into account to come up with better predictions about the future. Now there are thumb rules for using quantitative forecasting and that is:

1. Data for at least 2 seasons is needed (Data from Jan’17 — Dec’19 is needed to make a better prediction for Jan’20).
2. The prediction window should be the same as that of data (monthly forecasting model cannot be used to predict weekly or daily forecast).

#business-forecasting #retail-analytics #strategic-planning #business-analytics #forecasting #data analytic