Corporate Information Factory (CIF) Resources by Bill Inmon, Inmon Data Systems

> home > view content

Closed-Loop/Open-Loop Architecture

By W. H. Inmon

For many years the computer profession and business have formed a partnership that has operated under what can be termed an open-loop architecture. But with recent advances in data warehouse technology and the possibilities of the Internet, there is the possibility of what can be termed a closed-loop architecture for the marriage of business and computers. With a closed-loop business/computer architecture new business opportunities and possibilities arise that were never before possible.

An open-loop architecture is one where the business decisions made by management have an impact in the marketplace and where the impact of that decision is measured only indirectly by the corporation’s computer systems. As an example of an open-loop architecture, a bank vice president decides to raise interest rates. The results of the decision to raise rates are felt by the loan officers and those applying for loans and are measured indirectly as new loans come in or fail to come in. As another example of an open-loop business/computer architecture, suppose an insurance executive decides to offer a new kind of insurance. The market nods its approval or disapproval by buying or not buying the new insurance offering. By measuring the response in the open-loop architecture, the marketplace response to the insurance executive’s decision is quantified. As yet another example of an open-loop business/computer architecture, a manufacturer decides to improve its quality process. Over time the market signals either its approval or disapproval by purchasing or not purchasing the product. In every case of an open-loop business/computer architecture, the impact of the corporate decision is felt indirectly through the actions taken by the marketplace and is measured indirectly by the computer systems that run the corporation.

Such interactions between the decision process and the measurement of the effectiveness of the decisions made can be called an open-loop decision environment. In an open-loop decision environment there is no direct relationship between a corporate decision and a consequence, although there certainly is an indirect relationship.

Over the years business has operated almost exclusively in an open-loop decision environment. Prior to data warehousing it has not been possible to employ closed-loop business/computer architectures.

One of the challenges in an open-loop business/computer architecture is that of measuring the consequences of the decisions made in the corporation. One of the difficulties in measuring the results of an open-loop decision environment is that there is a lag of time between the moment the corporate decision has been made and the moment when the marketplace puts its verdict on the decision. It may be a long time before the marketplace responds to the decisions made by management. Another difficulty of measuring the results of corporate open-loop decisions is that in an open-loop environment, there may be factors other than the decision made by corporate management affecting the results achieved in the marketplace. While it is true that management’s actions have an effect on the marketplace, other factors also have an effect. Trying to measure precisely the effect of management’s decisions is an imprecise science at best. Trying to quantify the impact of management’s decisions in an open-loop environment is difficult, in every case. Trying to quantify the effect of open-loop corporate decision-making is an art much more than it is a science.

But with the advent of data warehouse technology coupled with the Internet, it is possible to start to measure the impact of corporate decisions in a very direct manner with what can be termed a closed-loop decision environment.

A closed-loop decision environment is one in which the impact of decisions can be measured very precisely. In a closed-loop decision environment management makes a decision, the computer system singles out exactly to whom the decision applies, and a direct action is taken as a result of the decision. The response to the action can then be measured on a case-by-case basis. The results of the management decision can be measured quickly and directly.

As an example of a closed-loop decision environment, suppose management selects all customers who have shown an interest in James Bond movies. The data about such an interest is stored in a data base and is based on information gleaned from click stream analysis. Now that the customers have been selected, they are offered a new James Bond T-shirt at a discount price. The results of the promotion can be accurately and quickly calculated. Measurement is easy to do and quantification is easy.

Or, as another example of a closed-loop business/computer architecture, suppose a department store looks to their data warehouse to tell them who has made a purchase of over $500 in the past year. Once the purchase information is collected, each of the customers who have made a purchase over $500 is offered a personalized credit card. In such a manner management can measure the effect of the decision quickly and very accurately.

With a closed-loop decision environment there is the possibility of having very quick and very accurate feedback on the decision process. The ability to make this assessment is very valuable.

But there is another very important aspect to the closed-loop decision process and that aspect is that the customer can be approached directly. Unlike open-loop decision processes where the customer is approached indirectly, with closed-loop decision-making, customers are named and numbered.

The technology underlying closed-loop decision-making requires that masses of data be handled in an organized manner and that there be a mechanism for directly interacting with the customer. Data warehousing satisfies the first requirement, and the Internet satisfies the second requirement. It is only with the intersection of data warehousing and the Internet that closed-loop decision-making can be done.

The differences between closed-loop decision-making and open-loop decision-making opens the door for an analysis to the kind of sales that are being made. Open-loop decision-making is for mass marketing companies. Companies that really don’t know who their customers are employ open-loop decision-making. As an example of a company who does not know who its direct consumers are, consider a soft drink company such as Coca-Cola or a tobacco company such as R.J. Reynolds. Coca-Cola and R.J. Reynolds sell to distributors. It is the distributors who know who the customers are. And in the case of vending machines and fast food outlets, the distributors merely know where and when to find the consumers, not who they actually are. So Coca-Cola and R.J. Reynolds engage in open-loop decision-making and sales.

Now consider a very different kind of company. Consider a bank or a telephone company. A bank or a telephone company must know its consumer, at least as far as retail sales are concerned. Now the bank or the telephone company may not know much about the consumer, but at least they know the name, address, and some other basic information. Unlike Coca-Cola or R.J. Reynolds, the bank and the telephone company do know who their consumer is. It is those kinds of companies who need to know how to engage in closed-loop marketing. Banks and telephone companies have the opportunity to be direct with their customers because they do know who the direct consumer is. Banks and telephone companies have the opportunity of directly measuring the results of their decisions.

But there is yet another kind of company that has opportunity at their doorstep; that company is one that today is a mass sales company that can turn into an individual sales company. Take for example a bookseller. Ten years ago how were books sold? You walked into B Dalton and you looked at the books they had for sale. You found one you liked and you bought it. Did B Dalton know who you were? Possibly they knew you through your credit card, but even then they didn’t know much about you. So back then when B Dalton made a marketing promotion, how did they do it? Through open-loop decision techniques. But along came Amazon.com. Amazon.com showed the world how to get to know your customers. Does Amazon.com know who their customers are? Yes. Does Amazon.com know what their customers are likely to like? Yes. Can Amazon.com get to the marketplace quickly and with measurable results? You bet they can. And how does Amazon.com do its thing? Through closed-loop decision and sales systems, backed up by a data base or two and the Internet.

With closed-loop marketing systems there is a difference in emphasis. With closed-loop decision systems the competition is measured on the basis of competition for a consumer’s dollar. With closed-loop decision systems the sales person is concerned with how much money a consumer spends on the goods offered by the closed-loop company. For example, from Amazon.com’s perspective, they are concerned with dollar competition for each of their customers with money spent on entertainment, research, cars, and even food. Every dollar spent elsewhere means less dollars for Amazon.com. Therefore Amazon.com is in competition with everyone.

This notion of who the competition is with closed-loop decision-making is in sharp contrast with the notion of competition with open-loop companies. Ask Coca-Cola who their competition is and they will tell you it is Pepsi-Cola or maybe Anheuser Busch. Or ask R.J. Reynolds who their competition is and they will tell you Skoal or Silk Cut. Open-loop decision makers look at competition on a product basis, not on a dollar basis, as do the closed-loop decision and sales companies.

One effect of the advances made in data warehouse and the Internet is that companies can now change from being an open-loop decision company to being a closed-loop decision company. For example, B Dalton can start to engage in both indirect and direct sales. This gives B Dalton the best of both worlds.

There is a cost to a closed-loop architecture. That cost is the cost of infrastructure required for the Web environment and the data warehouse environment. In a way, the Web and the data warehouse become the admission ticket for entry into the world of closed-loop processing.