The accounting software decision-making process is also important because the strengths and weaknesses of various packages differ. For instance, one package might include the standardized chart of accounts you wanted, but does not support the custom-designed invoicing feature you also need. You want a package that satisfies your major needs. It is less clear how to make the tradeoffs among packages.

A final reason for the importance of choosing among packages is cost. Modern accounting packages are remarkably inexpensive to purchase, but deceptively costly to install and operate. The costs of training, vendor support, data conversion, and backups are almost guaranteed to dwarf the software's initial purchase price. The software decision could also affect future hardware purchases.

The high stakes in purchasing accounting software make a careful, measured approach to software selection critical. Decision support tools can assist in this process.

Preliminary Steps

It is helpful to think of the software selection decision as a two-fold process. Your first step is to decide which features are most important. In the second step, you select the software that best matches these needs.

The features that you must have are usually dictated by the nature of your organization and can generally be determined by analyzing your existing accounting system. For instance, if you sell both products and services, you'll require software that can bill for both products and services. Similarly, if you require departmentalized accounting, you must have software that supports this. One caution here is not to trust someone else about vital features. For example, if you must have network capability or departmentalized accounting, check the software yourself or ask the vendor for references and contact them to verify features.

In addition to features you must have, you can also develop a secondary list of features you would like to have that are more a matter of preference and convenience. Examples include good user interfaces, speed, and vendor location. Advertisements are a good place to discover features that may be of interest.

While there are literally hundreds of features associated with any accounting software package, many are so common that they need not be part of the selection process. For example, most packages enable you to add accounts receivable customers "on the fly." Instead, you should focus on those features that are most useful to your own specific business. For instance, in a medical office with 50 doctors, you may need to track expenses for each doctor. Thus, a package that lets you create custom subaccounts is preferred over one that makes you specify every account in the chart of accounts.

Once you have made your lists, you will want to narrow your field of choices to five or so packages. The specialized accounting decision software discussed next can do this for you. Otherwise, you will need to learn about what accounting software is available and which packages have the features you need. For help, you can turn to trade journals which regularly evaluate accounting software for small and medium-sized businesses, trade advertising literature, demonstration disks, or free training sessions offered by vendors.

The likelihood is high that you will feel more comfortable with a final choice if you use some decision tools to help in this process. Three tools that can help you are: 1) decision support software developed expressly for accounting purchases, 2) generalized decision support packages, and 3) custom-developed models on spreadsheets.

Decision Support Software for Accounting Purchases

Probably the most direct decision support tools for choosing accounting software are the packages designed expressly for this process. Two examples are Requirements Analyst and Accounting Decision Maker, both of which include a data base of software features, a list of software packages, and a decision model to make a selection.

Requirements Analyst, which comes with the Guide to Accounting Software for Microcomputers, leads you through a two-stage decision process. In the first stage, you narrow the number of acceptable packages with a "features analysis" that identifies the features you must have, and an "elements analysis" that identifies desirable features you would like to have. There is no formal weighing scheme per se, but you can assign a "1" to features you must have and a value from "1" to "10" for features you would like to have. There is a total of 800 features, grouped by category, in the features data base--a daunting number for the casual user.

Once you have identified important features, you can proceed to evaluate the accounting packages in the system according to your criteria. Each package in the data base has been pre-evaluated by Computer Training Services as to whether or not it incorporates each feature. Finally, you can generate reports (Exhibit 1) to guide your software selection. Included are reports detailing the percent of Needs and Likes met by each accounting package and a listing of the Features/Elements met by up to four packages.

The Accounting Decision Maker is software that accompanies the Accounting library. This package includes several books and forms in addition to the decision support software. When using the Accounting Decision Maker, you first identify those features that are important to you, and assign them weights ranging from "0" to "9". (A "9" indicates that a feature is critical.) There is a list of about 900 processing features in the system's data base. Weighted averages are then calculated for each of the 100 accounting packages in the data base and final scores computed.

EXHIBIT1

SAMPLEOUTPUTFROMTHEREQUIREMENTSANALYST

PercentageofNEEDSandLIKESMet

BRANDNEEDLIKE

OPEN82%56%

MAC81%52%

CYMA80%59%

ASS78%70%

BPI77%15%

MAS68%45%

SOL63%55%

PLAT57%75%

EP56%44%

LIBR50%82%

SBT44%25%

GP37%60%

AM29%30%

Thisreportshowshoweachsystemscoredaccordingtothe

specificcriteriaentered

foryourcompany.Youmayevaluateasmanyorasfewsystemsas

youwish.

The output from the Accounting Decision Maker is a report showing the total score for each package, and a second score shows the percentage of package matches with features independent of weights. (You can reduce the package set by filtering certain ones.) Another report shows the highly rated features that are absent for the vendors with the top scores. As with the Requirements Analyst, when you purchase the package, you are also purchasing the right to buy updates.

Generalized Decision Support Packages

The weighing schemes of the two accounting software packages may fail to TABULAR DATA OMITTED capture the subtle tradeoffs in features that can make the difference between a system that is superior and one that is merely satisfactory. For this reason, you may prefer to use a generalized decision support package as an alternate decision support tool.

One example is Expert Choice, a microcomputer-based, interactive decision support system incorporating the analytic hierarchy process (AHP) developed by Thomas Saaty. The AHP approach views the decision problem as a hierarchy of needs, whose variables and decision alternatives are weighted to reach a decision. Since this is generally the nature of selecting accounting software, the process lends itself well to a hierarchical methodology.

To use Expert Choice, you first input the variables and choices that enable the system to build the model hierarchy. For the problem at hand, the top level in the hierarchy is the objective of choosing a specific accounting package. The next level contains features or attributes that are important in reaching a conclusion--for example, the package features of speed, ease of use, vendor support, number of customers supported, and quality of the chart of accounts.

Before recommending a specific package, Expert Choice requires you to compare criteria pairwise to determine their relative importance. Thus you must decide if "Speed" is more important than "Ease of Use," and if so, how much more important. You will then compare "Speed" with "Vendor Support," "Custom Financial Statements," and so forth, until you have assessed all pairs of features.

Once you have compared the features in the second level of the hierarchy, you next compare the variables in the third level. This means that you now compare Package A with Packages B and C with respect to "Speed," "Ease of Use," and so forth. For each comparison, you must not only decide which software package is preferable, but how much more preferable. Preference choices continue until all packages are compared against one another for each of the variables in the hierarchy.

After the comparisons have been made, the decision model in Expert Choice can synthesize your information and recommend a best package based upon these specifications. However, as with the other models, you should not rely entirely on this recommendation without further thought. The model is only as good your input. If the final recommendation is different from what you expected, it may be that you omitted some important criterion, or your answers were inaccurate or inconsistent. For instance, if you prefer package A to B and B to C for a specific feature, you should prefer A to C. Expert Choice calculates the amount of inconsistency in your judgments, and you can rerun the model with different answers to analyze the sensitivity of your choices.

While the comparison process is somewhat tedious, the advantage is that it forces you to make careful judgements because you weigh each variable against the others. You can limit the number of comparisons by either prioritizing features and including only the most important ones or by using two levels of the hierarchy, designating some features as criteria and others as sub-criteria. Another advantage is that the hierarchical structure of the decision process forces you to think about the problem in a logical, structured fashion.

Expert Choice is just one example of a generalized decision support system that can be adapted to the accounting software selection decision. Information about other such decision tools is included in the table at the end of this article.

Spreadsheets and Weighted Scoring

The most flexible decision support tool is probably a spreadsheet. With it you can construct any type of decision model you like. As suggested in Exhibit 2, your spreadsheet model is likely to list desirable features in separate rows, and use separate columns for software choices. Like the other models already described, you can examine hundreds of items, but limiting the number of features and packages keeps things simple. You may also find it useful to list evaluation criteria by accounting module. For example, in an accounts receivable module, you might list the number of customers supported, the ability to accommodate temporary customers, and the ability to prepare customized invoices.

Once you have identified needed features, you should next assign weights that indicate their relative importance, for example, the values between "1" and "5" illustrated in column (A) of Exhibit 2. Of course, these can all be the same value, but it is also highly likely that they will differ for any particular installation.

The best software package is the one that best incorporates the important features you have listed. Some sort of scoring device is necessary to determine this such as a ranking scheme that assigns weights to each feature in your list. Although the scoring weights for each feature can be on any scale, a suggestion is to use numeric values between "1" and "3". This range eliminates the need to make fine distinctions between packages and keeps things simple. Note that if you do not use the same scale to score each feature (e.g., do not use the values of "1" to "3" consistently), you are changing the relative weights you assigned earlier to these features.

Your last step is to compute a total weighted score for each accounting package. In Exhibit 2, this requires you to multiply each feature's weight times the package score for that feature and add the column. The package with the highest total is the recommended package - i.e., Package B (with a final score of 59) in Exhibit 2.

Some of the advantages of using a spreadsheet to help you reach a purchasing decision are:

* You can use existing spreadsheet software for this activity;

* Your spreadsheet model is highly visual and understandable; and

* You can use a customized weighing scheme that closely reflects the importance of each feature in your criteria list.

You can use your spreadsheet's graphics options to visually compare feature rankings. This approach also allows you to specify exactly which software features are most important, rank and re-rank vendors according to your own weighing scheme(s), and experiment quickly to examine the sensitivity of your answers to any change in criteria.

Using a weighted scoring approach on a spreadsheet to evaluate accounting software also has some drawbacks. For one thing, it requires effort gathering the data, inputting values into your computer, and designing the decision model itself. Unlike the software designed expressly for evaluating accounting software, you must also develop your own list of features. Finally, as with any decision support tool, you must remember that the model is designed to support decisions, not make them.

The best approach for selecting accounting software is largely a matter of needs and objectives. If you are recommending software to others, for example, packages such as the Requirements Analyst and the Accounting Decision Maker are valuable tools to have as references. Alternatively, if you are choosing a package for your own business, you may prefer a spreadsheet or generic decision support approach. These options better allow you to concentrate on your specific business needs and also spare you the trouble of wading through several hundred features before identifying the few that are really important to you.

The results of the decision support software or spreadsheet are not absolute answers. Quantifying decisions is a method of assisting in making judgments for which there are no right or wrong answers. The final decision must still be made by humans, not the computer.

It's All Relative

A final note of caution: you may ultimately pay more for a decision support package than for the accounting software it recommends. Do not be put off by this. The costs of installing an accounting package, operating it, training employees, converting data, and maintaining software and equipment are likely to far exceed the initial cost of a decision support package. More importantly, buying the wrong package is likely to cost much more.