By Donald F.
Cox
and Robert E.
Good
Recently the marketing vice president of a company whose sales volume is $350 million asked, "How should we go about developing a marketing information system? I don't mean one that will keep track of orders and shipments, but a system giving our marketing managers information that will help them make better decisions about pricing, advertising, promotion, product policy, sales force effort, and so forth."
He asked the question of us because, since early 1966, we have been studying the attempts of 15 major U S. corporations to develop a sophisticated marketing information system, or MIS. We have talked with executives at companies such as Chemstrand, Coca-Cola, General Electric, General Foods, IBM, Lever Brothers, Pillsbury, Schenley, and Westinghouse.
Although this field is relatively new, most of the technical aspects of developing an MIS are no longer an obstacle. Nevertheless, few companies are very far along in taking advantage of an approach which, its users agree, has great potential.
In this article we will attempt to provide some guidelines which might help answer the inquiring marketing vice president - and others with similar questions. First we will present a brief review of some of the characteristics and advantages of a sophisticated MIS, and of the current "state of the art." Then we will identify some of the key decisions which must be made by top management in the MIS development process. In each case we will present a distillation of the experience of the companies studied as an aid in making these critical management decisions.
What It Is & Can Do
An MIS may be defined as a set of procedures and methods for the regular, planned collection, analysis, and presentation of information for use in making marketing decisions. This of course is a step beyond logistics systems, which handle inventory control, orders, and so forth.
AUTHORS' NOTE: This is a partial report of a study of the development of marketing information systems which is being supported by the Division of Research of the Harvard Business School.
It is desirable first to differentiate between the two major components of such systems - support systems and operating systems. Support systems include those activities required to generate and manipulate data - i.e., market research and other data gathering, programming, and data processing. Operating systems are those that use the data as an aid to planning and controlling marketing activities.
This article is concerned mainly with the development of three types of marketing operating systems - those designed for control, for planning, and for basic research. In EXHIBIT I we summarize some of the applications and probable benefits of each type of system (assuming increasing degrees of sophistication) and present examples of systems now operating. The following are examples of marketing systems we have observed, with some of the advantages the companies claim for them.
1. Control Systems
These provide
continuous monitoring (sometimes through exception reporting) and rapid
spotting of trends, problems, and marketing opportunities. They allow better anticipation of problems,
more detailed and comprehensive review of performance against plans, and
greater speed of respon se. For instance:
- IBM's Data Processing Division has developed an MIS which district
sales managers can interrogate through a time-sharing computer terminal located
in an executive's office. A manager punches a typewriter-like keyboard and
receives an immediate print-out of information such as:
- Sales (or rentals) to date -
broken down by product code, type of customer, and branch making the sale.
- Sales in relation to goals.
- Combinations of information
which relate to sales, customer classifications, product codes, and so forth.
The data are current
to within three or four days, allowing the manager to keep up to date on marketing
problems and opportunities and on progress in relation to goals.
-Schenley has installed the so-called SIMR (Schenley Instant Market
Reports) system which allows key executives to retrieve (via video display desk
consoles and printers) current and past sales and inventory figures for any
brand and package size (or combination) for each of 400 distributors. SIMR
furnishes information in less than one second after a query, compared with many
minutes or even hours under its former computer and manual system. Furthermore,
since the computer does the calculations, managers have great flexibility and
near-instant speed of response in making many types of comparisons of sales and
inventory positions, such as:
- How a brand is doing in any
size or in all sizes in any market or in all markets.
- How a distributor is doing with
a particular brand.
- How a bottle size is doing by
distributor, state, or region.
- How a market is doing by month
or has done since the end of the previous fiscal year.
We can get answers
literally while we are still formulating the questions," states Bernard
Goldberg, president of Schenley's marketing subsidiary. "Needed
information is available so quickly that it helps us think."
2. Planning Systems
These furnish, in convenient form, information the marketing executive requires for planning marketing and sales programs. At least three major consumer goods producers, for example, are developing "data books" for product managers. The books bring together the basic information a product manager needs to formulate annual marketing plans and to "replan" during the course of the year. Putting the information into one book, rather than in a welter of reports, not only saves time, but it also enables all product managers in a group or division to base their plans on the same data. Consequently, their superiors are able to review comparable information quickly when considering the plans for approval.
At a more sophisticated level, planning systems allow simulation of the effects of alternate plans so that the manager can make a better decision. For instance:
- Pillsbury's system enables marketing managers to obtain sales
forecasts for each of 39 sales branches, supported by varying levels of trade
promotion. The marketing manager asks the question, "What will sales be in
each branch if we spend x dollars on trade promotions in comparison with .75 x
dollars and with 1.2 x dollars?" Pillsbury does not claim that the system
is perfect - it is obviously no better than the assumptions on which the
simulation is based - but it has had a surprisingly good "batting
average" in accuracy. It has great value to marketing managers because it
allows them to look at alternate plans in each of the 39 sales branches; this
was never feasible before.
- A large pharmaceutical company has developed an even more complex
model. The company has programmed an artificial panel representing the nation's
population of doctors. Every week the company simulates each doctor's
prescription decision for every patient he "sees." (Commercial
research services are available which provide information on the incidence of
symptoms of illness and the "patient mix" of the various medical
specialists.) The doctor considers the symptoms "presented" by each
patient and decides whether to prescribe a drug and, if so, which type and
brand. His decision is based on factors such as his experience with the drug,
current attitudes, exposure to the advertising of various brands, exposure to
detail men, and word-of-mouth information from other doctors. The simulation
even includes a "forgetting routine" which causes a doctor to forget
from time to time some of the information he has acquired.
While the company does not disclose how the simulation model is being used, it certainly is capable of generating extremely sophisticated marketing planning. For example, marketing managers can test the effects on share of market and sales of variations in amount, type, and timing of advertising and simultaneously test the effects of variations in frequency of detail men's calls. On a broader basis, the system can be used to screen a number of alternative marketing programs to select the most promising ones to be actually test marketed.
Perhaps the ultimate in sophistication is a marketing planning system which reviews alternatives, then actually makes decisions and takes action. Thus, several large retailing organizations have developed systems that review sales trends and inventories and then place orders for merchandise.
The most advanced unit of this type we have seen is not a marketing system; rather, it buys and sells securities in a stock brokerage house. Still in the future are marketing systems that decide the amount and timing of advertising and price promotions in each of several dozen sales districts.
3. Basic Research Systems
These systems are used
to develop and test sophisticated decision rules and cause-and-effect
hypotheses which should improve ability to assess effects of actions and permit
greater learning from experience. For instance:
- A large consumer goods company is developing an MIS which, among
other things, stores in computer memory the characteristics of each
advertisement run (color versus black and white, nature of illustration, amount
of copy, and so forth) and readership and attitude change scores for each ad.
The purpose is to be able to relate ad characteristics to effectiveness
measurements under different conditions and with different types of consumers
by systematically studying "experience."
- Most companies find it difficult to relate advertising to sales
because there are so many important "uncontrollable" variables which
are nearly impossible to take into account in an unsophisticated MIS. One large
consumer goods producer has developed an MIS which for the first time allows
the company to collect, store, and retrieve advertising, sales, and other
marketing data at a level of detail which makes possible much better controlled
studies of the relationship of advertising to sales.
Current Progress
The examples which we have presented probably represent the most sophisticated types of MIS now in existence. While we have not surveyed the 500 largest corporations in the country, we have screened more than 50 companies and have reviewed more than 100 current articles on information systems. As far as we have been able to determine, the current state of the art is something like this:
- Very few companies have
developed advanced systems, and not all of these are in operation. Some might
even best be classified as subsystems, since they relate to only a portion of
the marketing decisions made.
- Some companies, perhaps 15, are
actively upgrading their systems to a high level. Of these, about half seem to
be progressing well; the others have been much less successful.
- Many other companies are
contemplating plans to develop sophisticated systems.
The reasons why marketing systems have not developed to the same extent as, say, production, logistics, or financial systems are not "technical." Marketing research technology (data gathering), computer technology (data handling), and analytical procedures (e.g., mathematical model building) are all sufficiently advanced to permit companies to build effective marketing systems.
Although insufficient time has elapsed since the installation of most advanced marketing information systems to allow a precise assessment of benefits, the users of sophisticated systems with whom we have talked are virtually all very enthusiastic about their systems, even though many see room for improvement.
Developing the System
Because many of the
technical problems of developing sophisticated support systems have been solved
and many users are gratified over the results, why are there so few advanced
marketing information systems in operation? And why have some companies
succeeded more than others in realizing the potentials of the MIS?
One characteristic of
the more successful companies is striking. In every case, at least some members
of top management have seen the promise of the technique and have viewed its
development as a top management responsibility. They have devoted a great deal
of time, thought, and effort to guiding (and sometimes) actually protecting the
development process Unfortunately, it is widely believed that the job of
building an MIS can be turned over to a technical staff group. This has not
proved to be the case. Information systems are not merely technical appendages
(developed by technical people) that are easily meshed with most existing
marketing planning and control systems.
The best way to show
why participation of top management is necessary is to pose five key questions
which must be answered in the process of instituting a sophisticated MIS. In
our opinion, each is a management question:
- How should we organize to
develop a better MIS?
- How sophisticated should our
marketing systems be?
- What development strategy
should we follow - do we attempt to build a "total" system in one
move, or in stages?
- What should be the major
characteristics ("macro specifications") of our system?
- How much should we spend on
developing and operating an MIS?
While the field is too new to permit comprehensive and conclusive statements about all its aspects, we can present some guidelines and working hypotheses that are worthy of management's consideration.
Readying the Organization
The starting point in organizing for MIS development is not the establishment of a marketing systems group. The starting point is a review and appraisal of the entire marketing organization and of the policies that direct it. As James Peterson, vice president grocery products marketing of The Pillsbury Company, pointed out to us:
"We realized we couldn't develop a marketing control system until we had clearly and sharply defined the responsibilities of our marketing managers. If the system was to measure their performance against plans, we had to specify precisely what each man was accountable for.
Some companies, for instance, have failed to decide whether a product manager is accountable for unit sales and market share, for sales revenue, for marketing profit, or for net profit. Until responsibilities and spheres of activity are clearly defined, if is virtually impossible to build a marketing control system. In fact, specification of who is accountable for what automatically determines many of the control system's characteristics.
Management must next decide how to organize MIS development activities. Our observations show that this is a much more complex problem than might be assumed. Sophisticated systems require the coordinated efforts of many departments and individuals, including:
- Top management.
- Marketing management, brand
management.
- Sales management.
- New products groups.
- Market research personnel.
- Control and finance
departments.
- Systems analysts and designers.
- Operations researchers,
statisticians, and model builders.
- Programmers.
- Computer equipment experts and
suppliers.
The contribution of each group of course depends on its specialized talents and interests in the system. Programmers cannot define managers' information needs, and managers usually cannot program. No one person knows enough to accomplish all phases of MIS development.
Furthermore, sophisticated systems do not fall into a company's traditional data handling domains, such as the market research department or the accounting and control department, because an essential feature of a good MIS is that it integrates and correlates marketing and financial data.
Many companies we have observed have not really come to grips with the difficult problem of providing the organizational arrangements and leadership necessary for successful MIS development. They have not answered the question of who is responsible for MIS design, planning, and development. Why is there a leadership vacuum? Partly because top management does not fully appreciate the requirements and implications of the MIS, and partly because it has an understandable reluctance to disturb entrenched and powerful departments.
The approaches which have been tried in an attempt to solve the problems of organization and leadership can be characterized as:
- "Clean piece of
paper" approach.
- Committee approach.
- Low-level approach.
- Information
"coordinator" approach.
'Clean Piece of Paper' Approach. This involves drawing a new organization chart. The argument goes that the financial and accounting departments and market research departments have developed