Business Model Pattern Execution

A System Dynamics Application

Mar

ıa Camila Romero, Mario S

anchez and Jorge Villalobos

Systems and Computing Engineering, Universidad de los Andes, Bogot

a, Colombia

Keywords:

Business Model, Patterns, System Dynamics, Execution, Constructors.

Abstract:

The dynamic aspects of a business model are key to understanding the behavior of the business and the con-

sequences of any change. In spite of the multiple approaches to describe business models, most of them

emphasize static elements and leave the dynamic ones to intuition which in turn, diminishes the overall un-

derstanding of the model. Among the approaches that explicitly present dynamic elements, we ﬁnd business

model patterns which provide a visual representation of the dynamic behind the business by portraying infor-

mation, value and money ﬂows. Although the representation delivers adequate insights on the dynamics, it

is possible to enhance them by executing the patterns. To visualize the dynamic behavior and the effects of

changes and time in any component, we present an approach based on system dynamics.

1 INTRODUCTION

The study of business models is fundamental to

understand the enterprise’s behaviour and the ef-

fect of changes. Since businesses are complex sys-

tems that can only be analyzed by identifying parts

and relations among them, one must acknowledge

the dynamic behaviour and implement mechanisms

that support complexity management. The business

model becomes relevant in this scenario as it provides

the necessary description to understand the business.

Since it has become a topic of interest among re-

searchers, multiple approaches and deﬁnitions have

emerged.

One of these approaches is based on business

model patterns (Romero et al., 2016). In it, a busi-

ness model is deﬁned in terms of four core processes:

Transformation, Supply, Delivery and Monetization.

The last three can be described with patterns, as they

are support processes that do not differ signiﬁcantly

between enterprises. Each pattern is built in terms of

participants, activities and ﬂows. The last are key to

represent the dynamic behaviour of the business, as

it it possible to establish how value, information and

money are ﬂowing through the business’ different el-

ements. Furthermore, as the patterns are presented in

zones, they can be assembled depending on the nature

of the business.

Although the ﬂow representation gives a ﬁrst view

on the dynamic aspects of the business, it is neces-

sary to enhance their visualization to recognize the

effects that are unleashed as they establish relations

among components, and time goes by. This not only

provides a better understanding of the business, but

also allows to foresee the possible consequences of

any change. For doing so, one should execute the

patterns but, given that the current representation is

not executable it is necessary to modify it. This pa-

per presents a system dynamics application in which

business model patterns are expressed in terms of sys-

tem dynamics, and executed to visualize the different

relations and the possible effects of any change. By

providing an initial conﬁguration of a business model,

and deﬁning the rules that regulate each ﬂow, it is pos-

sible to execute it and describe the behaviour of the

business and the dependencies that determine it.

2 BUSINESS MODELS

As there is no universal deﬁnition on business models

the concept is not very clear therefore, one needs to

understand both what a model and what a business is.

A model is an abstraction of a reality; a representa-

tion that allows the analysis of a system without deal-

ing with the system itself (Selic, 2003). The business

on the other hand is often used as a synonym for en-

terprise. Enterprises are systems made up of interre-

lated resources that act together with the environment

to attain a particular goal (Giachetti, 2010). They may

440

Romero, M., Sánchez, M. and Villalobos, J.

Business Model Pattern Execution - A System Dynamics Application.

DOI: 10.5220/0006369704400447

In Proceedings of the 19th International Conference on Enterprise Information Systems (ICEIS 2017) - Volume 3, pages 440-447

ISBN: 978-989-758-249-3

range from companies to small stores that, in spite of

their differences, are characterized by a high level of

complexity. The business on the other hand, is tra-

ditionally perceived as a dimension of the enterprise.

Regarded as the function of an enterprise, it explains

what the enterprise does (Hoogervorst, 2009).

2.1 Business Dynamic Behavior

In spite of being a part of an enterprise, businesses

too are complex systems. This, not only because of

the amount of elements and relations that make it up,

but also, because of the complex dependency network

that is established as they inﬂuence each other over

time. This network determines a dynamic behaviour

that is responsible for the multiple effects that a single

change can inﬂict in the business. Understanding this

behaviour is essential for studying the possible con-

sequences of any alteration however, static business

model approaches are not enough for doing so.

Since some dependencies emerge from other re-

lations or are subject to time, they simply can not

be portrayed in static terms. Nonetheless, many ap-

proaches leave the dynamic behaviour to the mod-

eler’s intuition as the associated cost of the model is

higher. Therefore, it is necessary to ﬁnd mechanisms

that diminish the cost and encourage dynamic models.

The cost of the model may be evidenced in the con-

struction process or in the analysis, therefore, tools

that automate analysis or constructions processes are

highly valued as they save time and effort.

2.2 Business Model Patterns

The work presented in this paper is based on business

model patterns (Romero et al., 2016). These patterns

are built based on a proposed business model deﬁni-

tion: the way in which an organization gathers sup-

plies, transforms them into a product or service, de-

livers and monetizes it. This deﬁnition acknowledges

four processes: Supply, Transformation, Delivery and

Monetization. The transformation process is unique

for every enterprise but the remaining three can be

described in terms of patterns. All the patterns are

presented as zones that can be arranged depending

on the business characteristics. This allows to con-

sider them as modular constructors that contain the

description of a process and that, depending on its po-

sition on the complete model, will deﬁne the overall

behaviour. Each pattern contains participants, activ-

ities, gateways, processors and ﬂows. The last ones

are essential to portray the dynamic aspect of the busi-

ness.

3 SYSTEM DYNAMICS ON

PATTERNS

Although patterns can be used to analyze the busi-

ness’ dynamics, they suppose a higher cost. So, to fa-

cilitate their application one should establish a mecha-

nism to either fasten the building process, or simplify

the analysis. In order to reduce the cost associated

to the analysis one should ﬁnd means to facilitate ei-

ther the process of obtaining results, or the process

of interpreting them. One useful approach is System

Dynamics. With it, one is able to analyze the model

for a desired period of time based on an initial conﬁg-

uration. Furthermore, considering the structure of the

approach and the building blocks that must be used, it

is possible to maintain a relation with the visual rep-

resentation of the patterns.

3.1 System Dynamics

The complexity of systems emerges not only because

of the amount of components, but because of the re-

lations that arise with time. Although it is possible

to identify certain cause-effect relations, the inherent

feedback of the complex systems leads to other ef-

fects that can not be foreseen from a linear perspec-

tive. (Sterman, 2001). System Dynamics is a well

know approach to analyze complex systems and their

non linear behaviour based on stocks and ﬂows mod-

eling. Through stocks, one is able to represent ac-

cumulations (of materials, money, people...) that can

be measured in time. Flows on the other hand, are

the product of movement through time. Flows carry

what stocks accumulate and depending on their type,

they will increase or decrease the stock. The relation

between stocks and ﬂows is established through in-

formation links (Kirkwood, 1998). In order to create

a system dynamics models, one should identify accu-

mulations, ﬂows and, through the use of information

links, establish the equations that regulate said ﬂows.

3.2 Pattern Equivalence

Since business model patterns are presented in a static

way, they must be translated to an executable repre-

sentation. To do so, equivalences between the sys-

tem dynamics components and the pattern’s must be

found. Since system dynamics considers stocks, ﬂows

and information links, and patterns include zones,

ﬂows, gateways, processors participants and activi-

ties, there is not a one to one equivalence. Further-

more, deﬁning the corresponding representation also

depends on the way in which system dynamics will

Business Model Pattern Execution - A System Dynamics Application

441

be applied. In our case, to build the model and execute

it we used the software iThink by isee Systems.

3.2.1 Flows

The conversion of the pattern model begun by deﬁn-

ing ﬂow equivalences. This was based on the pur-

pose of the system dynamic model and in particu-

lar, in what was going to be analyzed. Since patterns

consider three ﬂow types, depending on the type that

would be monitored, the equivalence between stocks

and ﬂows was be established. For matters of this

research, the value ﬂow was given priority and so,

stocks and ﬂows were associated to it.

Information ﬂows were given the equivalence of

information links since there was an interest in visu-

alizing the movement throughout the different com-

ponents, but not in the particular values. However, if

an information ﬂow triggers another information ﬂow,

the result of the interaction should be modeled with a

converter (Auxiliary Variable) in order to be stored.

For instance, if a participant generates a request to get

a value from another participant in order to make a

calculation, the result of the latter needs to be stored.

This considering that the result will be used to trigger

other ﬂows. Converters should also be used when the

information ﬂow is associated to a status notiﬁcation.

Finally, money ﬂows were monitored with an Ex-

cel ﬁle that is updated as the model is executed. This,

since there was a need to keep trace of the accumu-

lation of money, and it allowed a better analysis of

ﬁnancial components. Table 1 presents the equiva-

lences between pattern and system dynamics ﬂows.

3.2.2 Participants and Activities

The equivalences of participants and activities were

granted with the stocks and ﬂows. As there is a con-

cern for the system’s overall behaviour, we did not

include activities as part of the equivalences although

they can be associated to the sources and sinks of the

ﬂows, that most if not all the time, are associated to

the accumulation.

Participants on the other hand, were represented

either by stocks or converters. The last ones being de-

ﬁned as auxiliary variables that keep trace of certain

values. The decision between both representations

depends on the participant and whether it associates

with a value ﬂow at some point. If the participant

generates or receives value then it is represented with

a stock; if it only receives money or information, then

it is represented with a converter (auxiliary variable).

Table 2 presents said equivalences.

3.2.3 Processors and Gateways

Considering the graphical notation of iThink and the

clear differences between the different ﬂow types,

processors were not used in the equivalence. Gate-

ways on the other hand, were already included in the

ﬂow representation of system dynamics ﬂow.

3.2.4 Zones

As patterns are enclosed in a zone, the iThink model

was build in a similar way. By drawing a frontier

(square) around each system dynamic pattern that

was modeled, it was possible to maintain a zone-like

equivalence. It is expected that the iThink zone con-

tains the system dynamic components relevant to the

portrayed pattern.

3.2.5 Pattern Conversion

The general conversion from patterns to iThink mod-

els is presented in Table 3. First off, one must identify

the number of zones in the pattern and portray them

with the equivalence. Next, participants are differ-

entiated between those that receive or trigger a value

ﬂow, and those that do not. In the general business

model pattern, the three participants are related to

value ﬂows and therefore they all are presented with

a stock.We then proceeded to deﬁne the value ﬂows.

According to the pattern there are two value ﬂows that

connect participant 3 and 1, and participant 1 and 2.

Consequently, the ﬂows in iThink should connect the

stocks in the same way. Finally, we represented the

information ﬂows with converters and ﬂows. To do

so, we followed the order established in the pattern.

In this case, we have a ﬁrst ﬂow that comes from an-

other zone, so, to represent and store its value one

should deﬁne a converter in the triggering zone and

in the receiving zone, both connected by an informa-

tion link. There is also an information ﬂow (generated

by the previous one) that triggers another information

ﬂow. In this case they deﬁne a decision and so, they

should be connected to a converter that generates the

fourth ﬂow.

The previous process was followed to convert all

deﬁned business model patterns.Table 4 presents the

conversion of the Deliver Engineer to Order Product

pattern. As it is possible to evidence the number of

iThink zones is correspondent to the number of zones

in the pattern. In this case it is important to note that

Logistics is presented as a converter as it does not in-

volve value ﬂows. Furthermore, as there are two ﬂows

that arrive/leave other zones, the correspondent stocks

of the zones are portrayed,

ICEIS 2017 - 19th International Conference on Enterprise Information Systems

442

Table 1: Flow Equivalences.

Information Flow iThink Component

Value Flow iThink Component

Table 2: Participant Equivalences.

Participant iThink Component

3.3 Study Case: Editorial

In order to test the equivalences and dynamic model,

we used a study case based on an editorial. Edito-

rial de los Alpes offers textbooks written by teach-

ers and professors for schools and universities in 12

cities.The edition of a book starts with the enrollment

of its author. To do so, a visitor stops by schools

and universities to promote the editorial services and

talk teachers and professors into the authors call. De-

pending on the visit success, a number of potential

authors will respond to the call and send the corre-

sponding manuscript. Considering the textbook de-

mand and the available manuscripts (from the catalog

and the new authors) certain books will be edited and

printed. Once the copies reach the warehouse, they

are delivered to the editorial ofﬁces in the remaining

11 cities. As the copies arrive, the city ofﬁce pays

for the corresponding order and proceeds to deliver

them to libraries and hypermarkets according to the

demand from their clients.

3.3.1 Pattern Model

To build the correspondent system dynamics model,

we ﬁrst had to build the complete business model in

terms of patterns. First, we deﬁned the zone conﬁg-

uration based on the processes that the editorial per-

forms. As portrayed in Figure 1, there are 7 zones.

2 supply zones (placed one on top of the other as

they are performed simultaneously) that account for

the materials needed to edit and print the books, and

the authors, 1 Transformation zone, 2 delivery zones

associated to intermediaries and the ﬁnal client, and 2

monetization zones (for the city ofﬁces and client).

With the zone conﬁguration we identiﬁed the pat-

terns. Figure 2 and Figure 3 present the two views of

the pattern model. In the case of Figure 2 it is possible

to see the supply zones with a source to stock pattern.

The transformation zone remains a black box with the

editorial performing the transformation process, and

one delivery zone (Deliver Stock Product). Figure 3

presents the city ofﬁce monetization zone with an as-

set sale pattern, and the delivery (Deliver Stock Prod-

uct) and monetization (Asset Sale) zones for the ﬁnal

client.

Figure 1: Editorial Zone Conﬁguration.

3.3.2 System Dynamics Model

Based on the pattern model and the previously de-

ﬁned equivalences we translated the model resulting

in the one presented in Figure 4. In order to build

it, a similar process to the pattern conversion was im-

plemented. First, zones were portrayed equivalently

to the pattern model, second, patterns were converted

with their initial equivalence maintaining the connec-

tions in the business pattern model. It is important to

note that the Editorial and Transportation participants

were modeled as conveyors (a special stock type) to

model the edition and transportation time.

4 PATTERN EXECUTION

With the executable model we proceeded to perform

its execution with two scenarios. The initial con-

ﬁguration run, that shows the behavior of the edito-

Business Model Pattern Execution - A System Dynamics Application

443

Table 3: General Pattern Conversion.

Pattern Composition iThink Equivalence

Participant 2

Participant 1

Participant 3

A3 A4

A5 A6

ZONE

Gateway

Activity

Time

End

Processor

Activity

Value

Cash

Information

Table 4: Sample Conversions.

Deliver Engineer to Order Product iThink Deliver Engineer to Order Product

Warehouse

Logistics

DELIVER

MONETIZATION

Gateway

Design

Approval

Product

Payment

Place Order

Dispatch Order

Deliver Order

Receive Payment

Time

End

Processor

Order

Client

Transportation

Status

TRANSFORMATION

Product

Status

Negotiation

Value

Cash

Information

rial with an initial parametrization, and the experi-

ment scenario in which two initial parameters were

changed in order to visualize the reaction of the edi-

torial.

4.1 Execution

The ﬁrst execution considered an initial conﬁguration

for the editorial, these parameters will be referred to

as normal conditions. The model was executed with

its initial conﬁguration in a 60 month run. Further-

more, the equations for gateways and converters were

deﬁned in terms of if-else business rules. In particular,

warehouses would dispatch materials depending on

the order, if the order was bigger than the warehouse

level, then the warehouse level would be dispatched,

if not, the order. Other equations considered demand

veriﬁcations that determined elements like the num-

ber of authors who answered the calls.

Table 5 presents part of the results of the ﬁrst run.

In it the following variables were measured: sales, ﬁ-

nal client (Accumulation of Sales), copies in edition

(Editorial) and in transit (Transportation, Car), raw

materials (Warehouse) and copies in store (City Ware-

house, Book Warehouse, Intermediary). All mea-

sures are done in terms of copies/month. For the raw

materials, the calculation was made considering the

amount of materials needed for a copy.

The obtained results show that under the ini-

tial conditions, the editorial’s capacity to respond to

client’s demands diminishes as delays in transporta-

tion avoid warehouses to receive an optimal number

of copies. Furthermore, as the intermediaries are the

ones that deal with sales directly, they present a dras-

tic reduction of their inventory as clients demands are

daily. In the end, this behavior can only be perceived

with time and that, in order to have a better perfor-

mance, the editorial should increase its capacities.

ICEIS 2017 - 19th International Conference on Enterprise Information Systems

444

Figure 2: Editorial Business Model View 1.

Figure 3: Editorial Business Model View 2.

The second execution considered two scenarios.

One in which the client demand was doubled and one

with a double edition capacity. For each scenario,

besides the modiﬁed parameter, every other condi-

tion remained as in the initial conﬁguration. Table

6 presents part of the results of both executions.

As it is possible to evidence, in the ﬁrst scenario

a double demand leads to sales doubling and an ear-

lier decrease in demand response capacity (month 10

and not in 40). The other behaviours remained similar

as they depended in transportation capacities were not

changed. For the second scenario inventories increase

Business Model Pattern Execution - A System Dynamics Application

445

Figure 4: Editorial Business System Dynamic Model.

Table 5: What if Analysis-Initial Behaviour.

Sales Edition

and in the intermediaries case, they tend to accumu-

late.

The execution of both scenarios contribute to the

further understanding of the editorials behaviour. It is

important to note that in spite of a regular behaviour

in several months, drastic changes may emerge as the

accumulation of delays will eventually affect perfor-

mance. Moreover, it is possible to evidence that edi-

tion and transportation capacities are key to the edito-

rial for they establish whether the business is able to

respond to clients demands or not.

5 RELATED WORK

The use of system dynamics in enterprise analysis is

presented in various works as the approach proves

useful to understand the overall behaviour of com-

plex systems. The simulation capacities of the ap-

proach have been used in researches regarding what

if analysis of enterprises (Sunkle et al., 2014), canvas

simulation (Romero et al., 2015) and resource alloca-

tion (MacDonald et al., 2003). In the ﬁrst case, what

if scenarios were tested for the intentional model of

the enterprise. Other works have also attempted to

understand and even forecast the nonlinear behaviour

proper of businesses. Such is the case of the work

presented in (Hoptroff, 1993) in which neural nets

were used to forecast and model relevant business el-

ements like cash ﬂows, stocks and sales. It is clear

that the complexity behind businesses has motivated

researches for understanding and managing it. Fur-

thermore, in order to guarantee an appropriate com-

prehension of the businesses, the effects of time in the

relations of the system must be acknowledged while

considering that static representations are not enough

to fully understand the business behaviour.

ICEIS 2017 - 19th International Conference on Enterprise Information Systems

446

Table 6: What if Analysis-Final Behaviour.

Sales (Double Demand) Edition (Double Edition Capacity)

6 CONCLUSIONS

This paper presented an approach to dynamic busi-

ness models by executing business model patterns

with system dynamics.By establishing equivalences,

a complete business pattern model based on a study

case was built and tested under two scenarios to

illustrate the scope of the approach. The results

clearly portray the effect of time on the relations be-

tween agents and the overall success of the enterprise,

and reﬂect the importance of evaluating the business

model under dynamic conditions for its complexity is

only revealed with time.

In spite of businesses being complex systems,

analysis like the one presented in this paper contribute

to complexity management. Although achieving a

complete understanding is not trivial, the fact that

from a static business model one can derive an exe-

cutable one that leads to valuable insights on the busi-

ness, proves to be useful. Approaches like system dy-

namics, which have systems as part of their core con-

cepts, provide an intuitive mechanism to construct-

ing executable models and easing their analysis. Fur-

thermore, if general equivalences are established, the

translation of static models to executable ones is ac-

celerated.

REFERENCES

Giachetti, R. (2010). Design of Enterprise Systems. CRC

Press.

Hoogervorst, J. (2009). Enterprise Governance and Enter-

prise Engineering. The Enterprise Engineering Series.

Springer Berlin Heidelberg.

Hoptroff, R. G. (1993). The principles and practice of time

series forecasting and business modelling using neural

nets. Neural Computing & Applications, 1(1):59–66.

Kirkwood, C. (1998). System dynamics methods–a quick

introduction. arizona state university.

MacDonald, B., Potter, J. M. M., and Jensen, K. O. (2003).

Long-term business modelling using system dynam-

ics. BT Technology Journal, 21(2):158–169.

Romero, M. C., S

anchez, M., and Villalobos, J. (2016).

Weaving business model patterns - understanding

business models. In Proceedings of the 18th Interna-

tional Conference on Enterprise Information Systems,

pages 496–505.

Romero, M. C., Villalobos, J., and S

anchez, M. (2015).

Simulating the business model canvas using system

dynamics. In 2015 10th Computing Colombian Con-

ference (10CCC), pages 527–534.

Selic, B. (2003). The pragmatics of model-driven develop-

ment. IEEE Softw., 20(5):19–25.

Sterman, J. D. (2001). System dynamics modeling: tools

for learning in a complex world. California manage-

ment review, 43(4):8–25.

Sunkle, S., Roychoudhury, S., Rathod, H., and Kulkarni, V.

(2014). Toward structured simulation of what-if anal-

yses for enterprise. In 2014 IEEE 18th International

Enterprise Distributed Object Computing Conference

Workshops and Demonstrations, pages 222–226.

Business Model Pattern Execution - A System Dynamics Application

447