A Knowledge Base Representing
Porter's Five Forces Model
Henk de Swaan Arons (email@example.com)
Philip Waalewijn (firstname.lastname@example.org)
Erasmus University Rotterdam
PO Box 1738, 3000 DR Rotterdam, The Netherlands
Abstract. Strategic Analysis and Planning is a field in which expertise and
experience are key factors. In order to decide on strategic matters such as
the competitive position of a company experts heavily lean on their ability
to reason with uncertain or incomplete knowledge, or in other words on
their experience and expertise. An important aspect is to assess a com-
pany's profit potential in the industry for which Porter's Competitive
Forces Model is by far the most widely used framework. This article fo-
cuses on the various aspects of designing and developing an expert system
representing Porter's Competitive Forces Model.
An important part of the Strategic Analysis and Planning concerns Porter's Competitive
Forces Model and it is interesting to see how this well-defined part can be modeled so that
the knowledge that it contains can be used in an expert system. This is the main subject of
Knowledge in the field of strategic analysis is either uncertain or incomplete. An expert in
the field generally will not have all data at his disposal. In particular, many data concerning
the environment of the enterprise, such as data of competitors and suppliers, will sometimes
be missing or is difficult to uncover and thus cannot be taken into account. But also data
from within the own enterprise is not always readily available. Whatever the reason for this
lack of data may be, the expert is expected to generate an analysis the best he can based on
data that is available. Naturally, the more relevant data he can use, the better the quality of
the analysis will be. In other words, if the input data is not sufficiently available, conclusions
drawn will be correspondingly less certain. This is one of the most striking characteristics of
an expert. He is able to come to conclusions based on a limited number of data that - in ad-
dition - may also be uncertain. For this purpose he will use his experience. It must also be
observed, however, that sometimes an analysis report must be generated in a very limited
period of time and then, whether available or not, the number of data used must inevitably be
limited. It is obvious that an expert system must also be able to deal with this type of heuris-
tic knowledge and come to conclusions that consequently will also be uncertain.
In the next section a brief survey is given of each of the fields of Strategic Analysis and
Expert Systems and how they could be related. In section 3 Porter's Competitive Forces
Model is briefly explained and it is discussed how various factors influence the profit poten-
tial of a company. Section 4 focuses on aspects such as the representation of this kind of
knowledge, the use of business rules, the inference process and reasoning under uncertainty.
Finally, in section 5 some conclusions are drawn.
2. Strategic Analysis and Expert Systems
Strategic analysis is a genuinely human task. Experience and expertise are generally consid-
ered indispensable for an analyst to assess the many data and circumstances that enter into an
enterprise’s position. Expert systems claim to perform this task provided that the experience
is modeled and processed properly.
In literature relatively little attention is paid thus far to the applicability of expert systems
in this specific field. Since the introduction of knowledge-based systems in marketing in the
late eighties, a growing interest emerged in expert systems, mostly as a part of decision sup-
port systems. The report Knowledge-Based Systems in Marketing  presents a thorough
study of the (potential) role of expert systems in this field and enumerates a number of do-
mains in which special purpose expert systems could be successfully applied. Unfortunately,
only a few of the systems developed so far have reached the operational stage. Most systems
have not (yet) passed the phase of prototype system.
Also more recent publications indicate that an increasing amount of expert system re-
search is being conducted for a diverse range of business activities . For example, a hy-
brid system for strategic marketing planning  that aims to provide a structured marketing
planning process, guides a user through this process, offers expert advice at key stated and
finally makes recommendations for users in setting objectives and strategies. It combines the
advantages of the expert system and decision support system technology in order to enhance
its effectiveness. Another expert system is one in strategic marketing with the objective of
helping marketing managers to analyze the position of their company relative to their com-
petitors, in a particular business or product area, and then suggesting ways in which this po-
sition might be improved . It is interesting to notice that this subject has a lot in common
with the commercially available product Business Insight  that is briefly discussed in sec-
3. Porter's Five Forces Model
A wide variety of schemes visualizing the strategic planning process are available. In es-
sence they usually consists of a series of steps (building blocks).
The analysis starts with defining the business and formulating a vision and then goes on
to assess the internal and external environment. The strategic planning process ends with the
financial budget and goes into a feedback loop (see figure 1).
Figure 1: Main steps in the strategic planning process
The essence of formulating a strategy is relating a company to its environment. Therefore the
analysis phase is crucial to the outcome of the total planning process. A major part of the
analysis phase is a diagnosis of the external environment. Several tools and techniques have
been developed to assist the planners in evaluating the external environment. Of particular
interest is the assessment of the profit potential in the industry.
Michael Porter’s Competitive Forces Model (commonly referred to as Porter's Five
Forces Model) is by far the most widely used framework for an assessment of the profit po-
tential in the industry. The collective strength of the so-called five forces (see figure 2) differ
from industry to industry.
Threat of new
Threat of substitute
products or services
Figure 2: Competitive Forces Model (Porter, 1980, p.4)
Each of those five forces is based on structural features (dimensions) which collectively im-
pact the profit potential. All five forces jointly determine the intensity of the industry com-
petition and profitability. The strongest forces become crucial from the point of view of
Our goal here is to address the most important dimensions of each of the five forces, for
they will be incorporated in the knowledge base (e.g. in the form of business rules) of the
proposed expert system. A number of important economic and technical characteristics are
critical to the strength of each competitive force. We will discuss them briefly.
Barriers to entry
These are the important structural components with an industry to limit or prohibit the en-
trance of new competitors. The major components are scale economies (advantage of experi-
ence, learning and volume), differentiation (brand image and loyalty), capital requirements
(new entrants will face a risk premium), switching cost involved by the customer, access to
distribution channels and cost disadvantages (patents, location, subsidies).
Rivalry among existing competitors
In most industries, especially when there are only a few major competitors, competition will
very closely match the offering of others. Aggressiveness will depend mainly on factors like
number of competitors, industry growth, high fixed costs, lack of differentiation, capacity
augmented in large increments, diversity in type of competitors and strategic importance of
the business unit.
These are products or solutions that basically perform the same function but are often based
on a different technology. Depending on the level of abstraction nearly everything can be a
substitution. In general the only factor that really matters is a shift in technology.
Power of buyers
Through their bargaining power buyers can force the competitors to lower their prices or
force higher quality or better service. The major factors which determine the bargaining
power are volume (relative to seller sales), does the product represent a major fraction of the
buyer’s costs or purchases, differentiation or standard product, switching costs, buyer profit-
ability (hence their price sensitivity), threat of backward integration, importance to the qual-
ity of the final product, and level of knowledge and information of the buyer of industry de-
mand, actual market prices and supplier cost.
Power of suppliers
Suppliers can exert their bargaining power over participants by threatening to raise prices or
reduce the quality. A supplier group is powerful if they are more concentrated than the in-
dustry they sell to, or if the customer group is not important for the suppliers, if the product
is an important input to the buyer’s business, or they have built up switching costs, or the
supplier group poses a threat of forward integration.
Through addressing these dimensions which make up the Five Forces we have outlined
factors which will be taken into account in our expert system. It will still be the expert’s in-
sight who will assert the value of impact of each individual variable. Another aspect is the
relative weight of each of the factors in the overall assessment.
4. The Expert System
A major representation technique are the so-called business rules, also known as rules of
thumb when uncertainty plays a role. Business rules commonly have the form:
if < a number of conditions is satisfied > then < a number of actions can be carried out >
Mostly actions are of the type: assign a new fact to the domain data base. Each business rule
represents a small, single chunk of knowledge and can easily be understood by anyone who
is familiar in the field. All of these business rules are combined by the inference engine that
can build up a reasoning chain.
In order to build an expert system that represents Porter's Competitive Forces Model it is
important to enumerate the factors that contribute to the final conclusion whether or not and
if so in which degree a company has a profit potential in its industry. In figure 2 the five
forces are enumerated: Buyers (bargaining power of buyers), Potential entrants (threat of
new entrants), Suppliers (bargaining power of suppliers), Substitutes (threat of substitute
products or services) and Competitors (rivalry among existing firms).
Each of these forces can be split up in smaller factors that build up to a conclusion about
the corresponding force. For example, as mentioned in the previous section, through their
bargaining power suppliers can raise prices or reduce quality. This is only possible when
they have the power to do so. And a supplier group is powerful if they are more concentrated
than the industry they sell to, or if the customer group is not important for the suppliers, or if
the product is an important input to the buyer's business. Factors that determine this bar-
gaining power are, among others, the level of the switching costs of materials, how impor-
tant is the supplier to the enterprise, and in reverse, how important the enterprise is to the
supplier, the degree in which the supplies are standardized, whether or not substitute materi-
als are readily available and whether or not the supplier poses minimal treat of forward inte-
gration. In Business Insight 5.0 (according to its tutorial, this tool is ‘A business factors
analysis tool that is used to gather knowledge and formulate strategies for business plan-
ning’) this complex of factors is taken together in a single business rule as follows:
Suppliers have moderate bargaining power (44) since
Cost to switch materials is high (10,6)
Supplier is very important to enterprise (10, 10)
Enterprise is not very important to supplier (20, 10)
Supplies are mostly standardized (70, 10)
Substitute materials are readily available (80,10)
Supplier poses minimal treat of forward integration (100,3)
The first of the two numbers related to each of the clauses indicates the rate on a scale rang-
ing from 0 to 100 in the case under investigation (e.g. in this case the fact that Substitute mate-
rials are readily available is rated 80) and this particular clause contributes to the conclusion with
a weight of 10 on a scale ranging from 0 to 10). The conclusion that Suppliers have moderate
bargaining power has been assigned a rating of 44. This rating is obtained by applying the for-
(10*6 + 10*10 + … + 100*3)/(6 + 10 + … + 3) ≅ 44
In order to show how the conclusion Suppliers have moderate bargaining power with a rating of 44
is used in other assertions the following assertion determining the strength of the production
operation is helpful (note that this specific clause has a weight 3):
The strength of the production operation is rated (47), since
Enterprise has poor economies of scale (25, 10)
Technology and production experience are unlikely to reduce costs (33, 3)
Production personnel skill level is average (43, 10)
Suppliers have moderate bargaining power (44, 3)
Enterprise has mediocre access to raw materials and personnel (54, 10)
Production manager is average (55, 10)
Enterprise may be cost competitive (60, 6)
Enterprise has fair control of material quality and production (62, 6)
The way Business Insight reasons with its assertions is a form of forward chaining: first the
user has to provide all necessary data based on which final conclusions can be drawn.
An obvious drawback of this approach is that it is only helpful when all data are readily
available. If data are lacking or incomplete Business Insight does not provide a way how to
achieve a valuable conclusion.
The same knowledge can also be expressed by a similar rule that needs to be processed
differently. For example, if all clauses of a rule are of more or less similar weight and each
clause can be qualified by 'good', 'average' or 'poor', then consequently the conclusion could
be either 'good', 'average' or 'poor'. According to some algorithm the choice could be made
which of the qualifications have to be chosen for the bargaining power .
the cost to switch materials is good and
the importance of the supplier to the enterprise is good, and
the importance of the enterprise is poor, and
the standardization of supplies is good, and
the availability of substitute materials Is average, and
supplier's requirement of forward integration is good
the bargaining power of the suppliers is average
From the corresponding assertion in Business Insight we learn that not all factors are of
equal weight and consequently they are not really suited to be taken together in one single
rule. In such a case the rule has to be split up in a number of other rules that meet this re-
quirement of equal weight.
A completely different way of formulating and handling business rules can be achieved
by introducing inexact reasoning. This approach circumvents a number of difficulties men-
tioned in the introduction (such as how to deal with incomplete data which is often a prob-
lem) but it also introduces a number of other complications. One of these is how to formulate
rules that incorporate uncertainty. An example of this kind of rules is the following:
the cost to switch materials is high, and
the importance of the supplier to the enterprise is high, and
the standardization of the supplies is average
the bargaining power of the suppliers is high (cf 0.8)
This rule states that the bargaining power is rated as high (with a certainty of 0.8 on a scale
ranging from -1 to +1) if all clauses are found to be true. This way of representing knowl-
edge and reasoning is based on the certainty factor model of Buchanan & Shortliffe . It
must be noted that each of the clauses of the rule are either the result of one or more other
rules or originates from some kind of data source (user, database, etc.). Examples of such
rules and how they can be combined to a reasoning chain can be found in . Similar rules
may come to the same conclusion although based on different observations such as the avail-
ability of materials, the importance of the enterprise to the supplier, etc. Other rules may
conclude that the bargaining power is either average or poor based on comparable observa-
tions. It will be clear that similar rules can be formulated for the other forces of Porter's
model. All these rules together form the knowledge base containing the knowledge of Por-
ter's model. The inference process can still be either forward or backward chaining.
This kind of knowledge representation can be implemented in expert system development
environments such as AionDS 7 or Exsys. Both of these environments support inexact rea-
soning. AionDS 7 also supports building object-oriented models which can be an advantage
as discussed in .
Porter's Competitive Forces Model is a well-defined part of strategic analysis and planning.
In this paper it has been discussed that there are several ways to represent the knowledge
contained in this model in an expert system. Each of the methods has its pros and cons but
the method of inexact reasoning seems to be best suited to deal with uncertain or incomplete
6. Literature and References
Wierenga B., Knowledge Based Systems in Marketing, Purpose, Performance, Perceptions and Per-
spectives, Management Report Series, no. 112, Erasmus University Rotterdam, Dept. of Business Man-
Wong, Bo K. and Monaco, John A., Expert System Applications in Business: A Review and Analysis of
the Literature (1977-1993), in: Information & Management [IFM], Vol. 29, Iss. 3, 1995.
Duan Yanqing and Burrell Phillip, A hybrid System for Strategic Marketing Planning, in: Marketing
Intelligence & Planning, Vol. 13 Iss. 11, 1995.
Moutinho Luiz, Curry Bruce and Davies Fiona, The COMSTRAT model: Development of an expert sys-
tem in strategic marketing, in: Journal of General Management, Vol. 19 Iss. 1, 1993.
User’s Manual Business Insight for Windows (An Expert System for Strategic Analysis), Business Re-
source Software, Austin, Texas, USA, 1993.
Swaan Arons H. de, Jellema M., An Object-Oriented Model of an Industrial Enterprise and its Envi-
ronment, in: Proc. European Simulation Multiconference 1997 (The Society for Computer Simulation
International), Istanbul, 1997.
Shortliffe E.H., Buchanan B.G., A model of inexact reasoning in medicine, in: Rule-based expert sys-
tems, pp. 233-262, Ed.: Buchanan B.G., Shortliffe E.H., Addison-Wesley Publishing Company, Inc.,
Amsterdam, The Netherlands, 1984.
Swaan Arons H. de, Waalewijn Ph., Strategic analysis modeled by heuristic knowledge, Proc. Fourth
World Conference on Expert Systems, Mexico City, 1998.