The Information Edge

 The web has far exceeded our expectations. It has become ubiquitous. Most firms have, or will have shortly, a presence on the Web and most are discussing e:commerce opportunities. However, there are other uses for the World Wide Web, though associated with e:commerce that have broader implications that just advertising, collecting orders and stockholder relations.

There are at least three types of novel productive Intra and Internet applications that have grown quickly over the past couple of years: (1) distance asynchronous learning, (2) marketing research, and (3) hubs of knowledge. These applications have potential to deliver improved competitive advantage through cost reduction and increased revenues.

Over the past two decades, corporations have become dispersed. Companies have not only become global, but have reached out to their customers and the value chain in such a way that the distinction between employees and agents and customers and partners have become blurred. Coupled with relatively high personnel turnover, in-company training and education have become increasingly important and difficult.

Utilizing the World Wide Web (with inexpensive custom CD's) is developing into an ideal vehicle for distance asynchronous training. The tools and procedures are quickly developing through the interests of the universities and supported by major software vendors such as Microsoft. Today it is feasible not only to provide "canned" slide shows, but to conduct interactive on-line discussions. Data suggests, that for some applications, there seems to be no tested difference between distance learning and in-person classes. This would suggest an enormous industrial opportunity.

We have just completed the analysis of our first on-line sophisticated marketing research survey. This study involved a number of pricing exercises that would have normally been executed either by personal interviews or a combination of phone and fax. The survey was completed in about half the time and half the fielding costs than expected with the alternate approaches. While there are still some problems in sampling using the Internet, for some studies it is the preferred research method. We believe that there is value in its application for business-to-business marketing and customer satisfaction studies.

It is insufficient for a web site to be interesting; it must compel the user to go to it. Developing web sites as a "hub of knowledge" drives the user to the site. It becomes the source of specific knowledge that engenders trust and ultimately sales. While this is not an easy task, it appears to be a highly effective web strategy.

While we expect some competitive advantage to be obtained by most participants in these applications, those who plan the process and effectively execute the development of the web sites early should obtain a disproportionate gain.

As discussed in our last newsletter, the purpose of market models and simulators is to forecast the impact of changes in strategy including pricing, new products and position. We had discussed using Microsoft EXCEL to construct user friendly decision support systems that incorporate market models and simulators. These models and simulators are often based on marketing research data.

The market pricing models discussed were designed to give forecasts for specific scenarios of competitive prices and potential product attributes. However, we live in an uncertain world. Competitive prices and actions are usually not known to a degree of certainty. Our clients want to know what is the overall risk of following a market strategy or pricing policy. By that they mean, what is the likelihood that they will meet the financial objectives that they have promised their management.

Market models and simulators do not capture all of the uncertainty in the world, but they do capture the key market reactions. By examining the outcomes of multiple scenarios, the sensitivity of price on earnings can be determined. This process can be expanded to include a large number of "randomly" selected scenarios. The result is a distribution of "probable" outcomes. This process is referred to as Monte Carlo risk analysis.

There are several add-on software packages designed to facilitate Monte Carlo analysis in spreadsheet. One of these packages, @Risk distributed by Palisade Corporation, was reviewed in an earlier newsletter (Vol. 1 No. 1, Summer 1996). We still believe that it is a useful approach for managers who wish to try the technique. However, it is unnecessary for the construction of decision support systems. Monte Carlo risk analysis can be done in EXCEL without any additional software. We have found that this approach has several advantages over using the add-in software:

1. It allows distribution of the decision support system to users without requiring that they have the add-in package.

2. It permits automation of the support system without the need for special training.

3. And, it allows for the use of bivariant risk distribution graphs that are not available in @Risk.

The market model is converted for Monte Carlo analysis by substituting the random number generating functions in EXCEL spreadsheets for the competitive prices or feature values. For example if we desired normally distributed prices for a competitive product with an average price of $5 and a standard deviation of $0.50 (10% standard variation) the EXCEL function would be:

Every time EXCEL recalculates, it would generate a different value for the price based on this function. Any number of scenarios are then generated using the Data/Table command. Graphs displaying the resulting distributions of earnings and share can then be plotted as shown below. These are automatically recalculated with changes in either the target price, new product features, or competitive prices.

Another approach is to determine the optimum price that would maximize the likelihood to meet a predetermined goal either in terms of earnings or share. This requires multiple Monte Carlo simulations. Once again this can be done directly in EXCEL by using the Data/Table command and assigning different values for price. The resulting plot is shown below. Please, note, however, that this computation may take a while.

While these approaches require more sophistication in their preparation and use than the standard decision support systems, they do answer the key management question.

While most of us remain behind the English language ramparts, there is a world that prefers to communicate in one of the scores of other major languages. That world has been slowly but inevitably making its presence known particularly over the World Wide Web.

But there is help on its way. Functional language translation, though not prefect, is available. For example, AltaVista provides on-line translation of web sites located by that search engine site from French, German, Italian, and Spanish to English and from English to those languages. The supplier of this service SystranÒ also provides packages for Chinese, Japanese, Portuguese, and Russian. In addition, there are other suppliers such as Universal Translator Deluxe that claims up to 33 language translations.

As with most Americans, I am unfortunately severely monolingual; I only speak and read English and American English at that. This, of course, makes my review of language translation less than stellar. However, I did find the Systran's utility (referred to as the "Babel Fish") very functional. With the exceptions of only a few names, I found the web sites understandable in the translation from German to English. My students, who were more than just fluent in the languages, were less forgiving of the roughness in the translations particularly with colloquial expres-sions. However, even they admitted that the translations would probably be adequate for technical or business applications.

As with all software, I expect these programs will improve. But, even in its present state, I feel comfortable using them to expand the range of my Internet searches as it truly becomes a World Wide Web.