Spreadsheet packages

News

21 February 1987

There is no doubting the value of such programs, and I would be the first to admit that it is essential for those engaged in the property sector to at least appreciate the basic spreadsheet concepts, in order to fully realise their potential. However, in my opinion spreadsheet packages, and application packages created using spreadsheets (sometimes called templates), do have serious limitations, which are frequently overlooked during system selection.

These shortcomings are most clearly illustrated in the case of property valuation which, on the face of it, appears to be an ideal application area for spreadsheets. Using the matrix of columns and rows a valuer is able to create valuation models with relative ease. (Note that models are created using cell references and hence no programming knowledge is required). Once developed the user can then enter different data to perform sensitivity testing.

How does this compare with, say, using a purpose-designed valuation program written in a high-level language (ie not created using a spreadsheet)? First, assuming it has been well written, the valuation program will be far more user-friendly than the spreadsheet, because the user will be guided through the program via a series of menus and prompts. At no stage should the user become lost in the system.

In contrast, users of spreadsheet models have to know how to find their way about the spreadsheet and it is all too easy to become confused.

Second, a well-written valuation program will contain data validation checks that prevent a user from entering spurious or inaccurate data. Such checks are extremely difficult, if not impossible, to build into a spreadsheet model. A further disadvantage of spreadsheets is that the models created are rigid and suit one given set of circumstances. Ideally, a computer program should provide a generalised, not a specific, solution, which may then be used for a wide variety of circumstances. The best valuation software does just that — users may adapt the given models to their particular requirements. For example, using a development appraisal program a user should be able to either enter basic information for a quick initial appraisal or comprehensive details of costs and values for a full valuation.

A further point is that the report-generation facilities of spreadsheets are somewhat limited, whereas with purpose-designed software reports may be tailored to suit the user’s requirements.

Finally, once a spreadsheet model has been created there is always that nagging doubt about its accuracy — surely the most important factor of all. In contrast, the generalised solution provided by the purpose-designed valuation system will have been rigorously tested.

Some may argue that the above problems can be overcome using a spreadsheet’s command language, but in reality such an approach is very tortuous compared with solutions developed methodically using a high-level language.

The spreadsheet package is certainly a useful tool for the property specialist — ideal for “one-off” problems — but for serious applications there is no substitute for the well-designed application package, written in a high-level language.

Since VisiCalc was introduced in 1978 the spreadsheet package has become firmly established as one of the standard pieces of microcomputer software. And today there are dozens of different systems available, all essentially based on the same concept. Of these systems there are two which dominate the market: Lotus 1-2-3 and Supercalc 4, both of which provide graphics and data storage capabilities in addition to the basic spreadsheet facilities. There is no doubting the value of such programs, and I would be the first to admit that it is essential for those engaged in the property sector to at least appreciate the basic spreadsheet concepts, in order to fully realise their potential. However, in my opinion spreadsheet packages, and application packages created using spreadsheets (sometimes called templates), do have serious limitations, which are frequently overlooked during system selection. These shortcomings are most clearly illustrated in the case of property valuation which, on the face of it, appears to be an ideal application area for spreadsheets. Using the matrix of columns and rows a valuer is able to create valuation models with relative ease. (Note that models are created using cell references and hence no programming knowledge is required). Once developed the user can then enter different data to perform sensitivity testing. How does this compare with, say, using a purpose-designed valuation program written in a high-level language (ie not created using a spreadsheet)? First, assuming it has been well written, the valuation program will be far more user-friendly than the spreadsheet, because the user will be guided through the program via a series of menus and prompts. At no stage should the user become lost in the system. In contrast, users of spreadsheet models have to know how to find their way about the spreadsheet and it is all too easy to become confused. Second, a well-written valuation program will contain data validation checks that prevent a user from entering spurious or inaccurate data. Such checks are extremely difficult, if not impossible, to build into a spreadsheet model. A further disadvantage of spreadsheets is that the models created are rigid and suit one given set of circumstances. Ideally, a computer program should provide a generalised, not a specific, solution, which may then be used for a wide variety of circumstances. The best valuation software does just that — users may adapt the given models to their particular requirements. For example, using a development appraisal program a user should be able to either enter basic information for a quick initial appraisal or comprehensive details of costs and values for a full valuation. A further point is that the report-generation facilities of spreadsheets are somewhat limited, whereas with purpose-designed software reports may be tailored to suit the user’s requirements. Finally, once a spreadsheet model has been created there is always that nagging doubt about its accuracy — surely the most important factor of all. In contrast, the generalised solution provided by the purpose-designed valuation system will have been rigorously tested. Some may argue that the above problems can be overcome using a spreadsheet’s command language, but in reality such an approach is very tortuous compared with solutions developed methodically using a high-level language. The spreadsheet package is certainly a useful tool for the property specialist — ideal for “one-off” problems — but for serious applications there is no substitute for the well-designed application package, written in a high-level language.