- Web Business Engineering: Steps to Planning a Successful Business Web Site
- What the Map Says About Using the Web
- Step 2: Model Activity Value
- Step 3: Diagnose Problems and Opportunities
- Step 4: Design Treatments
After creating the information activity map (Step 1) and the value model (Step 2), you are ready to diagnose the business process, looking for problems or opportunities that the Web can help eliminate or take advantage of.
Previously, we built an information activity map of a business process and a value modelspecifically, a cost modelbased on that map (refer back to Figure 7).
Next comes the fun part: using the model to determine problems or opportunities in the process that the Web can eliminate or take advantage of. If you have Microsoft Excel installed on your computer, you can interactively follow the discussion in this article by downloading the sample cost model created in the previous article.
Diagnosis: The Problem with "Common-Sense" Solutions
Your goal in diagnosis depends on your client's goals. Suppose that your client tells you that he or she wants to minimize the cost of the mailing process. The beauty of the cost model is that it allows us to systematically determine what factors drive this cost, such as the cost drivers. Without the cost model, you run the risk of suggesting "common-sense" solutions that have little or no effect.
For example, one common-sense solution to reducing costs in a process is to hire cheaper labor. A naiïve consultant may suggest reducing costs by hiring cheaper instructors at a wage of $25 per hour instead of $50. However, our model shows that cutting an instructor's wage (IWAGE) in half results in only a 1.15 percent reduction in mailing costs (see Figure 8).
The effect of hiring cheaper instructors on the mailing cost (Reprinted with permission from Flor, N. , Web Business Engineering, Addison-Wesley Longman)
Another common-sense solution to reducing process costs is by having people work twice as fast. In our model, this amounts to cutting all the time variables (PTIME, DTIME, CTIME, MTIME, STIME, RTIME, and WTIME) in half. However, our model shows only a 2.44 percent reduction in mailing costs when you do this (see Figure 9).
The effect forcing people to work twice as fast. (Reprinted with permission from Flor, N. , Web Business Engineering, Addison-Wesley Longman)
If common sense isn't helping, how can we determine what factors, if any, drive the cost of the process?
Diagnosis: Systematically Determining Costs
No techniques are guaranteed to identify cost drivers, but we can use several heuristics with our model to help us systematically identify them. One useful heuristic is the "one-half" trick. Using this heuristic, you reduce each factor in half and note its cost effects. Those factors whose reductions result in the biggest decrease in the overall cost are the likely cost drivers. The one-half heuristic suggests the following cost drivers: the number of sites that the packages are mailed to (NSITES), the mailing fee (MAILXFEE), the frequency of assignment mailings (TFREQ), the number of instructors (NINST), the number of weeks in a semester (WMINI), and the number of semesters in a year (NMINI). Cutting these factors in half results in cost reductions of 47.7 percent, 47.4 percent, 50 percent, 50 percent, 50 percent, and 50 percent, respectively (see Figure 10).
Applying the one-half heuristic (Reprinted with permission from Flor, N. , Web Business Engineering, Addison-Wesley Longman)
Now that you've identified cost drivers, you need to determine which drivers your client has control over.
Diagnosis: Determining the Variant Drivers
For organizational or cultural reasons, not all cost drivers can be eliminated. For example, there will always be four participating sites (NSITES), and instructors will always mail out two assignments per week (TFREQ). For similar reasons, WMINI and NMINI are not variable. The only factor that our client can really vary is the mailing fee. For example, the client can find a cheaper mailing serviceone that charges $3 per mailing instead of $20and reduce costs by 80.58 percent (not shown). Or, the client can eliminate paper mailings entirely by somehow using the Web. This corresponds to a zero mailing fee (MAILXFEE) and a 94.8 percent reduction in mailing costs (see Figure 11).
The effect of reducing mailing costs to zero (Reprinted with permission from Flor, N. , Web Business Engineering, Addison-Wesley Longman)
Note that our model doesn't really tell us how to use the Web to achieve this cost reduction. It merely suggests that any Web solution that drastically reduces or eliminates the mailing fee will result in a large reduction in costs. The important point, however, is that we aren't guessing what the problem isinstead, we're systematically determining it. Next, we examine how to design Web solutions, or treatments, that eliminate the problems that we uncover in our diagnosis.