Mova Business Forecaster: Description

Mova Business Forecaster is a data analysis tool for forecasting time series.

Mova Business Forecaster is a simple forecasting application that demonstrates the benefits of using external factors in a forecasting model.

The application uses historical Ford Motor vehicle unit sales. It uses a popular statistical model (SARIMAX) to predict the time series and forecast 24 months into the future. The model uses up to 14 user-selectable external factors (exogenous variables) that affect the accuracy of the forecast in different ways.

Easily select or deselect each external factor (exogenous variable) and see how it affects the accuracy of the forecast.

See how better forecasting methods can be used to improve your business planning and decisions

Why use a forecasting tool?

Many companies do not have sufficient data or do not have useful analysis tools for proper business forecasting:

• Spreadsheets often are too simple, contain errors, and are difficult to maintain.
• Intuition (gut feeling) is prone to bias, noise, subjectivity, emotion, experience, over-load, and accountability.
• Surveys and opinions take time and money, are subjective, are biased, are skewed, and can vary widely.
• Dispersed, siloed, outdated and incorrect data.
• Facts and sound analysis can provide more accuracy and certainty.

Why use external factors in a forecast?

Forecasting business metrics is an important activity for every company. Many companies use very limited forecasting methods, such as simple spreadsheets, simple historical analysis, or simple percentage changes.

Most companies do not include external variables in its forecasting models. Many people assume past sales cause future sales. Nothing could be further from the truth. Factors in the future drive (cause) future sales. These factors are constantly changing and impact future sales, independent of past sales.

This can lead to uniformed, limited or poor business decisions because managers do not have an accurate measure of where the business is and where it is going – or what is causing it. The future looks more different every day.

Including exogenous variables can improve forecast accuracy and confidence because the external factors actually drive (cause) the dependent variable. For example, hot weather increases ice cream sales. Or, higher interest rates reduces consumer spending.

Many time series forecast confidence intervals spread very wide in the future time period, which reduces the usefulness of the forecast. A simple best guess could be as good.

With more accurate forecasts, business managers and planners can make better plans and be more confident in their decisions.

Use external factors in your forecast methods to get more accurate forecasts and be more confident in your planning.

More accurate forecasting enables proactive business decisions and actions. Do not be reactive and wait for the problem or opportunity to show. Do not simply hope the past repeats itself. Use better forecasting to increase the gain or decrease the loss of the situation.

Select the Exogenous Variables to Improve Forecast Accuracy

The application’s SARIMAX model has 14 exogenous variables (external factors) that can be selected as a predictor of the motor vehicle sales. These factors affect motor vehicle sales in the following ways:

• Covid-19 Intervention: Job/work losses, reduced consumer confidence, work from home.
• Disposable Income: Affordability, cash, credit.
• Fed Funds Rate: Interest rates and borrowing costs.
• CPI New Vehicle: Vehicle prices.
• CPI Used Vehicle: Vehicle prices, compare with new vehicles.
• Employment Level: More people working, more jobs, more spending.
• Unemployment Level: Fewer people working, fewer jobs, less spending.
• Total Nonfarm Employees: Total number of people employed in non-agricultural sectors of the economy. More people, jobs, spending.
• Average Weekly Hours: Afford major purchases.
• Average Weekly Earnings: Afford major purchases.
• Retail Gasoline Price: Vehicle choice, miles driven, disposable income.
• Monthly Average Temperature: More shopping, seasonal, regional, vehicle choice.
• Monthly Total Precipitation: Less shopping, seasonal, regional, vehicle choice. We chose Chicago to model annual precipitation levels.
• U.S. Light Weight Vehicle Sales: Passenger cars and light trucks, including SUVs, pickups, and vans. National market sales.

The user can select between none and fourteen of the exogenous variables to be included in the model.