**2.2 KPI modeling**

### *2.2.1 KPI modeling issues*

The KPI modeling seems to be one of the most important actions closely related to KPI indicator processing and they have to indicate appropriate properties in order to be denoted as the good KPIs.

However, those properties may be seen as the first aspects concerned with KPI validation as well, while they night be postulated as follows [2].


However, the above-mentioned KPI indicator properties seem to be only one side concerned with KPI indicator modeling [15, 16].

**The second aspect** is closely related to KPI attributes postulated as KPI name, Type, Scale, Source, Owner, Threshold, and Hardness.

**The third aspect** is the performance indicator expression. It is "a mathematical statement over a performance indicator evaluated to a numerical, qualitative or Boolean value for a time point, for the organization, unit or agent. For example, P I27 ≤ 48h."

The **fourth aspect** concerned with KPI formalization is the performance indicator expression. It is "a mathematical statement over a performance indicator. The authors suggest specifying the required values of KPIs as constraints coming from goals. The authors claim that they integrate the performance view with the process, organization and agent-oriented views. However, there is no information about the process semantics used for modeling and no evidence about validation of the PI properties. In any case, the authors write about the process views of the real organizations, not about the abstract processes that are proposed [9, 15, 16].

#### *2.2.2 KPI modeling approaches and methods*

#### *2.2.2.1 MetricM method*

The method MetricM [18] "is built upon and extends an enterprise modelling approach to benefit from he reuse of modelling concepts to provide relevant organizational context, including business objectives, organizational roles and responsibilities." The method can be adapted to any enterprise modeling approach. The modeling language.

MetricML used in MetricM "adds essential concepts to modelling performance indicators and semantics to key modelling concepts." The concept Indicator is used to present a KPI.

*Business Process Linguistic Modeling: Theory and Practice Part I: BPLM Strategy Creator DOI: http://dx.doi.org/10.5772/intechopen.95096*

The MetricML Indicator metatype is used for modeling its relations to other indicator types, to reference object types representing organizational context and to goal types [2].

### *2.2.2.2 Attribute approach*

predictive models of future performance based on specified business objectives

The KPI modeling seems to be one of the most important actions closely related

However, those properties may be seen as the first aspects concerned with KPI

to KPI indicator processing and they have to indicate appropriate properties in

• KPI needs to be sensitive to changes of the business process state.

• KPI should be oriented to improvement, not to conformance to plans

However, the above-mentioned KPI indicator properties seem to be only one

**The second aspect** is closely related to KPI attributes postulated as KPI name,

**The third aspect** is the performance indicator expression. It is "a mathematical statement over a performance indicator evaluated to a numerical, qualitative or Boolean value for a time point, for the organization, unit or agent. For example,

The **fourth aspect** concerned with KPI formalization is the performance indicator expression. It is "a mathematical statement over a performance indicator. The authors suggest specifying the required values of KPIs as constraints coming from goals. The authors claim that they integrate the performance view with the process, organization and agent-oriented views. However, there is no information about the process semantics used for modeling and no evidence about validation of the PI properties. In any case, the authors write about the process views of the real organizations, not about the abstract processes that are proposed

The method MetricM [18] "is built upon and extends an enterprise modelling approach to benefit from he reuse of modelling concepts to provide relevant organizational context, including business objectives, organizational roles and responsibilities." The method can be adapted to any enterprise modeling approach. The

MetricML used in MetricM "adds essential concepts to modelling performance indicators and semantics to key modelling concepts." The concept Indicator is used

• KPI should be linear, (d) a KPI should be semantically reliable,

validation as well, while they night be postulated as follows [2].

and resource allocations [17].

order to be denoted as the good KPIs.

• KPI should be efficient,

P I27 ≤ 48h."

[9, 15, 16].

*2.2.2.1 MetricM method*

modeling language.

to present a KPI.

**162**

• KPI should be in a quantifiable form.

*Operations Management - Emerging Trend in the Digital Era*

side concerned with KPI indicator modeling [15, 16].

Type, Scale, Source, Owner, Threshold, and Hardness.

*2.2.2 KPI modeling approaches and methods*

**2.2 KPI modeling**

*2.2.1 KPI modeling issues*

An alternative "attribute" approach conceptualizes performance indicator as (meta-) attribute of metatypes (e. g "average throughput time" of a business process type or "average number of employees" of an organizational unit type). Alternative approach for KPI modeling in our method is used. MetricM uses declarative models. The model of underlying processes needed for validation of KPI properties are not used in MetricM. The two approaches, presented above, build upon ideas of many earlier approaches to KPI modeling. The general tendency is to postpone the validation of the KPI properties to the moment when the process model of the organization is ready.

#### *2.2.2.3 Semantics synchronous and asynchronous modeling*

However, the KPIs are defined at a different level of abstraction, namely at the tactical and strategic level, i.e. at the level of observable states of the system and the asynchronous modeling does not provide the right level of abstraction [15, 16].

The synchronous modeling semantics is based on the CSP parallel composition operator defined by Hoare [19]. The operator defines that an event from environment is accepted by the model if all processes of this model are able to accept it. Otherwise, the event is refused.

Although there were many applications of the CSP parallel composition operator in the architecture description languages [20] in programming languages [21] only after the extension of this operator for machines with data, made by McNeile [22] the operator became practical for business system modeling. The Protocol Modeling proposed in enables coping with complexity of business modeling. The reason is that the synchronous semantics decreases the data space of models.

#### *2.2.2.4 KPI indicator linguistic modeling approach*

This approach is based on existence of linguistic sets, while they represent KPI modeling static aspects. However, there are many relations among those linguistic sets as well, while they are quantified via PBPL Equation [10, 11, 23, 24]. This approach is discussed in Section 4 in more details.

#### **2.3 KPI indicator decomposition**

The KPI indicators are designed and closely related to core business processes implemented and operated at strategic management level and have a nature of so called initial and primary KPI indicators, which should be decomposed to secondary and tertiary KPI indicators. The secondary indicators are closely related to main BP management at tactic level and the tertiary KPI indicators are closely related to subordinated and elementary BP management at operational level. This approach to KPI indicator decomposition is discussed in Section 4. However, the KPI decomposition is closely related to business dashboard existence [13, 25, 26].

*A dashboard in business is a tool used to manage all the business information from a single point of access.* It helps managers and employees to keep track of the company's KPIs and utilizes business intelligence to help companies make data-driven decisions.

There are 4 general subtypes of dashboards: (a) Strategic - focused on long-term strategies and high-level metrics, (b) Operational - shows shorter time frames and operational processes. (c) Analytical - contains vast amounts of data created by analysts and (d) Tactical - used by mid-management to track performance.

**4.1 BPLM strategy creator – structure and functionality BPLM strategy**

*Business Process Linguistic Modeling: Theory and Practice Part I: BPLM Strategy Creator*

Any business is getting started by business mission statement and business objectives and adequate business process establishment. Those three categories create an integral part of any business strategy. However, before we determine a set of business quantitative and qualitative indicators, real possibilities should be known to apply our business results at an appropriate market area and collect initial information. Usually, the information is stored at different media and documents. However, we have to make a preliminary document content semantic analysis in order to gain a required information and this is an initial action, which should be done with the use of the proposed BPLM Strategy Creator. This type of the document semantic analysis enables providing the document categorization and show use which documents should create basis for processing of business strategy qualitative aspects. Furthermore, we are interested in those documents, which contain data closely related to business strategy quantitative aspects, which might be quantified via indicators denoted as key performance indicators (KPI indicators). However, they usually are not in that form and shape as we need. Therefore, we have to provide the second type of document analysis in order to extract required data – usually denoted as the initial data, which should inform us which products related to our business could be accepted by the market, in which quantity and quality and what about financial assets could be gained. This data type could create content of

Assetst]}. Because the data are of a linguist nature those

. This is only one side of the coin, while we to know what about

Costs (0)]} linguistic sets. The data represent the first BPLM Strategy crea-

Assetst *;* XTotfin

Costs ð Þ <sup>0</sup> *;* ½ � SAD ið Þ *;* <sup>j</sup> *;* ½ � HR ið Þ *;* <sup>j</sup> *;* ½ � TECH ið Þ *;* <sup>j</sup>

Costs ð Þ <sup>0</sup> ,

Assetst]}, contains data

Costs (0)]},

(1)

**creator – structure and functionality – qualitative view**

*4.1.1 Strategic management level*

*DOI: http://dx.doi.org/10.5772/intechopen.95096*

sets {[YTotfin

{[XTotmat

postulated

**165**

f g ½ � KPI 0ð Þ <sup>¼</sup> <sup>Y</sup>Totfin

to material assets<sup>1</sup>

Assetst]}, {[YTotmat

sets are denoted as linguistic sets. The linguistic set {[YTotfin

mentioned output products and they are stored within {[XTotfin

closely related to financial assets and the linguistic set contains data closely related

tor output, which is called the **basic output** as well, while the financial costs play a role of principal importance, but are not sufficient for production getting started. We have to know what about customers will buy our products, what about human resources with required theoretical knowledge and practical skills, and what about production technological devices and tools are needed, as well. This types of data

quantifies potential customers denoted as mainframe customers, the linguistic set {[HR (i, j)]} quantifies mainframe human resources and the {[TECH (i, j)]} linguistic set quantifies mainframe production technological devices and tools. A qualifier "mainframe" indicates that the linguistic set content is not specified in more details. When adding that linguist sets to the above-mentioned basic output we get the **initial BPLM** Strategy creator output, while formula (1) might be

investments (financial costs) are needed in order to pro produce the above-

are being stored in further linguistic sets. The linguistic set {[SAD (i, j)]}

Assetst , YTotmat

which represent so called the **total initial KPI** indicators.

XTotmat

<sup>1</sup> How many pieces of the actual output products could be produced.

**A strategic dashboard** is a reporting tool for monitoring the long-term company strategy with the help of critical success factors. They're usually complex in their creation, provide an enterprise-wide impact to a business and are mainly used by senior-level management [27–29].

**An analytical dashboard** is a type of dashboard that contains a vast amount of data created and used. They supply a business with a comprehensive overview of data, with middle management being a crucial part of its usage.

**A tactical dashboard** is utilized in the analysis and monitoring of processes conducted by mid-level management, emphasizing the analysis.

Then an organization effectively tracks the performance of a company's goal and delivers analytic recommendations for future strategies [30].
