• Use information you developed yourself. To be completely in control of your own analysis, you may limit source information to that developed within the department. The occasions when you can reasonably do this will be rather limited. Using only your data is practical when your analysis is limited to departmental factors.
• Compare your information to other sources. Many forms of information are easily verified by checking with a second source. Also, the existence of a second source could eliminate unnecessary research for your analysis.
• Review methods for reporting change from one period to another. Be extremely careful in developing any form of analysis involving reporting of changes in factors from one time period to another. There is a vast difference between reporting actual change and degree of change; some methods may distort the truth rather than represent it.
• Be fair. Avoid selecting only that body of information that supports your own bias. Because you are human, you are subject to human error. One form of error is to seek out information that supports the conclusion you have already drawn, or that you would like others to draw. You may do this unconsciously, even when struggling to remain as objective as possible. A good analyst will always remain open-minded to the possibility that basic assumptions are flawed. No one likes to admit this, especially if a number of hours have already been invested in preparing the analysis. But remember, it’s better to catch the error before making a request based on flawed data, than to proceed knowing your claim is wrong.
• Represent the other point of view to add credibility to your report. A decision maker will be impressed by a comprehensive and well-research report. The same decision-maker will be even more especially impressed when you go the extra step: playing devil’s advocate and pointing out arguments against what you’re recommending. This is a powerful technique for strengthening your argument. By presenting arguments against your ideas before anyone else has the chance to do so, you defuse the point, show why and how it’s flawed, and then direct attention back to the argument you really want to make: why your idea should be given approval.
• When results are interpretive, say so in your report. Never let it be assumed that the numbers you present are absolute. If you’re using estimates, disclose that fact in a way that there can be no mistake. And if the numbers were developed from other data in an interpretive form, say this clearly.

Historical Information
Historical information is all around you. Your department’s budget is one form of historical financial information. Some data will be non financial. For example, your employees have executed a specific number of transactions, worked varying numbers of hours, or used an office machine at an increasing rate over the past year. All these historical facts can be accumulated, organized, and used in requests, analysis, future budgets, and reports you generate.

However, remember: It’s a mistake to take too much comfort in consistency, especially at the expense of accuracy.
Historically based information may not serve your purpose completely. In many situations, an analysis may need to combine historical with other forms of data. Assuming that your analysis will be reviewed, perhaps even challenged, by others will help you to develop valid arguments and explanations.

This is the ultimate test of analysis, even when no one else will ever see it. Would you be able to support your conclusions if discussing them with someone else? How did you define and then collect your raw information? How was your analysis developed? Was it manipulated to create the outcome you wanted? Or did you test and question it to ensure that your study was fair and objective?

It’s a good idea to build a source file of recurring factors in your department. You probably need to develop your source file over a number of months. Correctly developed, this file can become a valuable collection of intelligence, to be used later in many ways.

Validating the Ratio
In many companies and departments today, invalid analysis is being cheerfully performed but never checked. Why? Because there is no means for finding out that the analysis is misguided. No one ever acts on discovered information coming from the analysis. If no one responds, what is the point of doing an analysis and discovering that something is wrong? A key point: The analysis is worth doing only if you and others are able to act in response to what you discover.

The response is where you discover whether or not your analysis has any merit. If your analysis shows a negative trend and you take action to correct it, you would expect to see results. However, if those results don’t occur, either the action you selected was ineffective, or the trend itself is not accurate.

If you are asked to perform an analysis, especially one that is part of a recurring routine, but no one ever acts on what it reveals, resist doing the task. It’s a waste of time. You might not have the power to simply refuse, but there are a number of steps you can take:
1. Eliminate the work itself. Suggest cancelling the routine. There is no value to creating important information no one uses.
2. Recommend the appropriate actions. Suggest actions others might take. Point out the problem: The information is important, but no one is acting on it. One way out is to suggest ways the problem can be solved.
3. Take over and do what’s needed. Volunteer to take the actions yourself. If no one else wants to do anything in response to what your analysis reveals, take on the job yourself.
4. Reduce the time commitment. Suggest completing the task quarterly instead of monthly. This at least reduces your wasted time by two-thirds. If you can’t completely eliminate a useless task, at least do it as little as possible.

After an analysis has been done several times, you will be able to audit your work. Ask yourself: Do the trends reveal what we’re looking for? Are the ratios dear and do they convey information to others? If not, how can they be changed or replaced? To validate the trend, check indicated actions against the trend itself. Do you see a change in direction as the result of changes you have prompted?

The validation process can now be undertaken. Did the changes you suggested reverse the emerging negative trend? Did the number of days required to collect outstanding accounts decline and stabi¬lize? If they did, it indicates that the ratio is valid, that it does give you valuable information.

You will notice, however, that validation can only occur if actions are taken in response to what the ratio shows. If there is no response, there can be no validation. A lack of response validates only one thing: that the analysis itself is a waste of time and will continue to be unless someone changes the situation.