To develop your research goals, outline the first page of a possible report on your project. Make sure that outline shows a possible recommended course of action, even if your end user has not asked for one. After all, the CI will be used by someone to make a decision. Begin by restating the needs statement as a conclusion and as an action statement. For example, if the question is "How can we catch up with competitor A in terms of costs of production?" you can, for example, restate it as follows:

Our competitor's most significant cost advantage is in its use of CAD-CAM [computer-assisted design-computer assisted manufacturing] systems in its manufacturing activities. [Remember, this is only hypothetical.] For our firm to catch up, we must change our manufacturing processes to incorporate the newest European CAD-CAM technology.

Following the action statement, list the kinds of key findings that might tend to support your recommendation. Following this example, you could write:

Our competitor has, and will continue to have, substantially lower costs than we do for manufacturing the same products. This is due to its CAD-CAM systems plus lower costs of raw materials.

The experience of manufacturers in other, similar industries indicates that CAD-CAM systems can be successfully adapted to our manufacturing operations within a short period of time, providing important cost savings.

Now, again transform these statements into questions. First, convert the conclusion you drew in Step 1 into the following:

• Does our competitor have lower costs?

• Why are its costs lower?

• For the two most important reasons for lower costs, can we duplicate or improve on these?

You are now drawing closer to a statement of the focus for your CI research. Knowing the research focus is crucial. It enables you to control and limit the data gathering efforts to those that have the greatest likelihood of responding to the general needs statement from your end user.

Also, once you know the focus of the research for the assignment, you will also be more likely to identify other data of potential interest to you and your end user, or leads to that data, even if it does not go into your final report. For example, here, if your research disclosed that a major consulting firm had just created a CAD-CAM systems consulting practice, you might want to save that information for future reference because you or your internal client might well be seeking such expertise after the CI project is over.

Now, repeat the conversion process in Step 3 for all of the draft key findings you developed in Step 2, like this:

• Does our competition have substantially lower costs than we do for manufacturing? If they do, will that continue?

• Are existing CAD-CAM systems sufficiently developed to be able to support our manufacturing techniques?

• What has been the experience of manufacturing firms in other, similar industries with the same CAD-CAM systems? Specifically, can it be successfully adapted to our manufacturing operations within a short period of time and will this provide a cost savings?

If you need to, you can repeat the process until you have created a series of smaller questions on which you can focus your CI research efforts. As you progress, you should find that, if any of your small assumptions are not correct, you can quickly adjust your focus while the research is underway.

Your CI research and collection strategy must reflect the unique needs of each assignment. Among the critical points to take into account in developing a research and collection strategy are the following:

• time constraints

• financial restrictions

• staffing limitations

• self-imposed constraints

Time is one of the most critical elements in limiting CI research. There are a wide variety of time constraints, some, or even all, of which may be applicable to a particular project. They can be most easily understood by asking, and then answering, the following questions:

• How long will data collection probably take?

• How much time do you have to conduct the project?

• How much of the total time is allocated to data collection?

• How long will it take to collect the raw data that you really need, that is, to answer the most important questions, but not all questions?

• Can you start your analysis work with only partial data (if so, when, or else what data do you require)?

• How much time is available for analysis?

• How much of your own time is available, and when?

In general, the key factor here is for you to determine when the CI is needed and then work backward. For instance, if you have three weeks to develop data on the market plans for new products of three key competitors, planning to go to a trade show that will be held in four weeks is obviously not a viable option. However, knowing about the trade show may cause you to try to change the assignment, by trying to get a later due date. For example, you might want to spend the time before the trade show researching marketing strategies and then attend the trade show or review the trade press coverage of the show to complete the research.

Similarly, if you are tracking a competitor over a prolonged period of time, using an article-clipping service may be a cost-effective way to collect background and general interest articles about the firm. However, if the analysis of the same competitor is just a one-time project or has a short time frame, you might do better using on-line commercial databases, even through their costs average $250 an hour. That is because these services provide you the data you need virtually immediately and permit you to go back in time, whereas a clipping service has delays in providing materials and generally only allows you to go forward. That is, you cannot ask the service to go back in time and add another topic to its monitoring.

Overall, you must appreciate (and, in turn, teach your end users) that time and money are always trade-offs in CI research. As a rough rule of thumb, cutting the time available for a project by half can be expected to at least double the costs of acquiring the same raw data. If time constraints are too critical, this can cause costs to escalate significantly. For example, your research strategy may identify four possible avenues of research. If you have sufficient time, you might logically pursue the one with the highest likelihood of success first. Then if that does not provide enough data, you would move to the second, and so on.

However, if you are put under severe time constraints, you might have to engage in parallel activity. That is, instead of proceeding sequentially through research options, you may have to start all four at once. You do not have the time to do otherwise. And the costs will rise accordingly.

The amount and timing of CI staff resources that you can put on a project at a particular point in time can be a very important constraint. Other types of staffing constraints include considerations such as how many people are available, what each of them can and cannot do, and when they have time available. Typically—if there is a typical case—a CI assignment requires fewer personnel during the planning and early collection phases than it does in the middle, when the data collection and review work begin to pile up. And at the end, bottlenecks will tend to develop as stacks of unanalyzed data accumulate.

Self-imposed constraints can vary widely. You may be comfortable with contacting your competitors directly. However, you may also find that you can no longer do this due to a change in company policy. This may force you to approach future projects from a different angle.

Another constraint may be a legitimate concern about allowing too many of your employees to learn about a CI project. That is, you or your end users may not want other employees, perhaps even other CI staff members, to be involved with the assignment at all. That can and has happened in cases involving a potential acquisition. This constraint can be so severe that you may have the choice of either abandoning all other projects to take this on yourself or turning it over to outside CI firms.