by: Murat Ergen, ForecastWare.com
Forecasting helps to obtain a good comprehension of the observed data and accurate future predictions based on this comprehension. The main goal, however, is to assist executives with decisive information they absolutely need to make the best possible decisions. Acting on these recommendations can significantly aid in solving problems, gaining new opportunities, and ultimately better business management. Examples of possible uses of forecasting are to reach corporate goals, to plan external economic and competitor impacts, to improve inventory management, to improve customer service, and to increase profit margins.
Key elements in forecasting are Team and Framework. In this article, characteristics of both are mentioned.
A successful forecasting system starts with forming a good team. Essential team members are Business Analyst, Data Analyst, Project Manager, Statistician, and Solution Architect. Roles and responsibilities each member contributes to organizational direction. Business Analyst makes sure credibility of the business information through careful research. Data Analyst automates historical file extracts, and optimizes the process time and accuracy. Project Manager has a unique position in forecasting projects, because of responsibility of asking details about the system and its sponsors. Statistician is responsible to find what is unknown, the forecasts, by studying the actual numbers and applying internal or external factors. Solution Architect works with system results as well as other available data to create a visualization tool for business leaders to make data driven decisions. With this team in place the following benefits could be expected; collecting business owner needs, and changes in the business, business approval of pre-published numbers, increased timeliness of business decisions, what-if scenarios implementation, short timeline delivery, easy to use custom solution interface, integration of major economic indicators and competitors market share, unnoticed real trends discovery, maximized profit margins opportunity, strategic business portfolio creation, on the spot tactical recommendations, centralized one version of truth, collaboration with other official numbers, automated step by step process, mitigation of known and unknown risks, and institutionalized ongoing maintenance.
The Framework for the forecasting system should include the following phases; Data Gathering, Information Gathering, Modeling, Reporting, and Recommendations. In the data gathering phase, historical data is extracted to be studied in order to gather historical data at the most granular level possible, input missing data points and attributes, structure hierarchy levels, and determine the business focus hierarchy level accuracy is tested. Information gathering phase follows what collected data is to note internal business circumstances, implement structural brake notifications, and detect anomalies and outliers. Modeling phase is where normalizing and transformation of data is taken in to account and studying the relationship between observed data and economic and competitor indicators. It this phase also models are estimated, validated, tested and selected to create forecasts. Reporting phase of the framework helps in both forecasts quality and applicable information selection such as static accuracy testing, dynamic best and worst performance analysis. Recommendation phase where Return On Investment are clearly observed because this phase tells business users where the potential goal gaps are, what resources to be added in business operation, and simply brings various departments to one table to list actionable items.
Some of the items need more emphasis;
- - Such as experienced statistician who has industry knowledge and built dynamic forecasting engines previously
- - The systems results needs to be reviewed regularly, this review includes business expected numbers and model numbers accuracy testing, what is working and what needs to be improved, model attributes
- - Constant Business and Forecasting Team communications
- - Selecting forecast engine tool for the corporate need and building a customized visualization tool for result comparisons such as actual to actual, goal to actual, forecast to goal
In conclusion, this article is about what is essential for a successful forecasting project. It is teamwork with applied framework. As long as these two are in place building, adjusting and maintenance of forecasts for the whole organization is an easy task. ForecastWare.com is built on this fundamental approach.