You have a spreadsheet full of figures, such as sales, costs and market data, and you suspect a few of them drive the result you care about. You can't tell which ones, and you need a forecast you can defend. Regression Analysis and Forecasting from Business Spreadsheets is an Excel template that does this work inside a workbook you already know how to open.
Quick facts
- Format: Excel template (multiple regression and forecasting)
- Platforms: Windows (Excel 97 and later) and Mac OS X (Excel for Mac)
- Price: USD 14.00
- Data limits: unlimited observations, up to 20 independent variables per run
Key features
Plain-language regression results, no statistics degree needed
Results are broken into sections with explanatory text beside the numbers. You read what the analysis means instead of decoding a table. The F-statistic significance is checked at 95% confidence, and the maximum confidence level is shown as well.
Per-variable statistics to weed out weak drivers
You get standard errors, t-statistics and p-values for each independent variable, alongside the overall multiple regression. You can drop variables with no significant relationship and rerun the analysis.
Checks for autocorrelation and multicollinearity
Tests for significance, autocorrelation and multicollinearity are calculated and explained. That lets you add and remove variables step by step toward a better equation. A chart plots actual against predicted values so you can see how well the fit holds.
Forecasts using the trend line that fits each variable
Each independent variable is compared under linear, 2nd polynomial, 3rd polynomial and exponential trends, with their relative strength shown. You can also leave the forecast input empty and supply your own external forecasts or assumptions. Set the number of periods, then export the output to a new workbook.
Input checking that cleans your data
Input data is checked for numeric values before processing. Non-numeric entries are cleaned and replaced in the input area, so you can see what changed. The first column holds the variable you want to predict and the following columns hold the drivers.
Who it's for
It suits finance staff, analysts, project planners and small business owners who build business plans or budgets in Excel and want to know which inputs matter. The design is generic, so it also handles time series and scientific data.
Pros and cons
- Pro: Results are explained in text.
- Pro: Several validity tests are built in.
- Pro: It runs in Excel, so there is no separate statistics package to learn.
- Pro: The price is low.
- Con: Each run is limited to 20 independent variables, so larger sets must be split across runs.
- Con: It needs Excel, so you can't use it without the spreadsheet program.
- Con: The forecasting options are trend lines (linear, polynomial, exponential) only.
FAQs
How many variables can I test?
Up to 20 independent variables per process, with unlimited observations. You can run separate processes to test more.
Do I need to know statistics?
No. Results are written out in explanations, so you can interpret and act on them without statistical training.
Can I use my own forecasts for the drivers?
Yes. Leave the independent variable forecast input empty and use external forecasts or alternative assumptions.
Can I take the forecast out of the template?
Yes. A button exports the forecast output to a new workbook for further analysis or presentation.
System requirements
- Windows: Excel 97 or later
- Mac OS X: Excel for Mac