Modeling the dynamic alignment of business and information systems via the lens of human-centered architecture evolution

ZHANG Mengmeng ,YI Shuanghui ,CHEN Honghui ,LUO Aimin ,LIU Junxian ,and ZHANG Xiaoxue

1.Science and Technology on Information Systems Engineering Laboratory,National University of Defense Technology,Changsha 410072,China;2.National Key Laboratory for Complex System Simulation,Military Academy of Sciences,Beijing 100080,China

Abstract:The complexity of business and information systems(IS) alignment is a growing concern for researchers and practitioners alike.The extant research on alignment architecture fails to consider the human viewpoint,which makes it difficult to embrace emergent complexity.This paper contributes to the extant literature in the following ways.First,we combine an enterprise architecture (EA) framework with a human viewpoint to address alignment issues in the architecture design phase;second,we describe a dynamic alignment model by developing a humancentered meta-model that explains first-and second-order changes and their effects on alignment evolution.This paper provides better support for the theoretical research and the practical application of dynamic alignment.

Keywords:business and information systems (IS) alignment,human viewpoint,enterprise architecture (EA),meta-model,dynamic alignment.

1.Introduction

The importance of business and information systems (IS)alignment has been well known and documented since the late 1970s [1−5].Over the past several decades,it has remained as one of the top-ranked concerns of business and information technology (IT) executives,contributing to initiatives such as“maximizing the return value from IT investments”,“providing direction and flexibility to react to change”,and“improving company performance”[6].Significant progress has been made in understanding and addressing alignment issues in recent years,mainly in streams of research on alignment measurement [7−9],enterprise evolution and management [10−12],and enterprise architecture (EA) research [13−15],as well as these topics’ relationships with other subjects [16−18].However,due to the emergent complexity and nonlinearity caused by social influence factors [19],dynamic alignment is still difficult to achieve.In the organizational evolution process,individual behaviors and human inertia are difficult to control,which hinders the achievement of sustainable alignment [20].

The methodology of EA is a beneficial way to achieve business and IS alignment.The purpose of the discipline of EA is to effectively align the strategies of enterprises with their processes and resources (business and IS).Many association studies have discussed various aspects of this discipline,such as EA design [13−15],EA evolution [10−12],and EA measurement [7−9].

Currently,many scholars have emphasized the importance of the human viewpoint in EA research [21−23].Different from other viewpoints in IS architecture,the human viewpoint is required to explicitly represent human factors and to document the unique implications that humans bring to system design.This viewpoint enables an understanding of the human role in systems and EAs,which provides a basis for stakeholder decisions by providing a structured linkage from the engineering community to the manpower,personnel,training,and human factor communities.It provides early coordination of task analysis efforts by both systems engineering and human factors teams.Currently,common EA frameworks fail to capture the human-centered design aspects necessary to ensure the effectiveness of human operated and maintained systems [21−23].Applying a human viewpoint to the architecture framework helps to plan the alignment architecture from the top down and to capture human behaviors from the bottom up.

Through the above explanation,to address human factors in the architecture design phase and to achieve continuous alignment,it is beneficial to consider the application of human viewpoints to the architecture framework.This paper introduces two contributions:first,we introduce a novel lens to combine architecture framework with the human viewpoint to capture human behaviors in alignment architecture;second,we describe the dynamic alignment of business and IS by developing a human-relevant meta-model and alignment evolution model,which explain the external changes and their influences on alignment evolution.

The remainder of this paper is structured as follows:Section 2 presents the related work;Section 3 mainly describes a human-relevant meta-model combining the North Atlantic Treaty Organization (NATO) human viewpoint with the meta-model of the Department of Defense Architecture FrameworkF (DoDAF);Section 4 develops an alignment evolution model with the above human-relevant meta-model and mainly explains the firstorder changes and second-order changes in the evolution model;and Section 5 discusses further studies,contributions and limitations of this paper.

2.Related work

To support our research,we separately explain related works from the extant alignment research,the alignment research with the EA method,and the human-related research on the architecture framework.

2.1 Extant alignment research

Throughout the research history of business-IS alignment,two distinct conceptualizations have been situated in different regions of complexity.The first is located in the region of order [24],where alignment exhibits high degrees of stability.A great deal of research has been conducted to establish such an alignment from multiple types[25,26],levels [6,27],and dimensions [28,29].Multiple alignment techniques have been recommended,such as business process management,EA,and service-oriented architecture.Although static and dynamic alignments are differentiated in this region,all of these techniques view alignment as a balanced [25],cause-effect deterministic logic [20].

The second conceptualization is situated in the region of emergent complexity [30].The alignment in this region moves away from equilibrium and exhibits features such as emergence,nonlinearity,self-organization,and coevolution [24].This kind of alignment goes beyond traditional alignment assumptions and challenges prior methods and models.A coevolutionary theory is deemed to be a superior way to address this research challenge[20].This theory expresses alignment as a coevolutionary process that reconciles top-down rational control and bottom-up emergent adaptation [20].Beyond the intended planning of alignment,the coevolutionary process aims to coordinate business and IS through continuous adaptation and learning [31].In our previous work,we classified the extant coevolutionary literature on alignment from dimensions including focus,framework,level,method,and phase [32].As a result,we found that it was still difficult to describe the dynamic nature of the alignment using the extant coevolutionary research.

According to the literature,alignment complexity mainly results from the social dimension,such as unintended deviations caused by human inertia,human errors,and lack of communication [20].Human behaviors are difficult to ascertain and control within the“dancing rugged”external environment and amid IT innovations[33].Capturing the human viewpoint is a key solution to enabling the pursuit of continuous alignment.

2.2 Alignment research with the EA method

EA is defined as the“fundamental organization of a system,embodied in its components,their relationships to each other and the environment,and the principles governing its design and evolution”[34].Enterprise architects seek to align enterprise processes and structures with their supporting IT systems so that enterprises can flourish in their environments [35].This alignment seems to be an instinct of EA development,and the evolution process from as-is EA to to-be EA can satisfy the dynamic requirements of the alignment.As Ross argued,EA is a payback of the alignment achievement [36].

It is widely acknowledged that intimate connections exist between alignment research and EA research[37,38].Many association studies have discussed various aspects of this discipline,such as EA design [13−15],EA evolution [10−12],and alignment measurement [7−9].In our previous research,we examined alignment research combined with the EA method using research questions,research methods,and techniques [39].A main limitation that results from the research examining this combination is the lack of human-relevant alignment management research.The discipline of EA is a top-level research method through top-down planning,which makes it difficult to control bottom-level human behaviors.Enhancing human-related exploration is important for future EA research.

2.3 Human viewpoint in architecture framework

The human viewpoint enables an understanding of the human role in EA.It provides a basis for stakeholder decisions by providing a structured linkage from the engineering community to the manpower,personnel,training,and human factor communities. An EA framework defines a common approach for development,presentation,and integration of architecture descriptions.The application of a framework enables stakeholders to contribute more effectively to building interoperable systems and to managing the associated complexities.Currently,most EA frameworks fail to capture the human-centered design aspects necessary to ensure the effectiveness of human-operated and maintained systems [21−23].For example,the open group architecture framework (TOGAF)[40] and DoDAF [41] describe the roles and objectives of performers but seldom explain human structures and human constraints,which limits them to capturing dynamic individual behaviors.

NATO provided a human viewpoint to model human resources and their applications in architecture evolution[21].The purpose of the NATO human view (HV) is to capture human requirements and the way that humans interact with other elements of a system.The models,descriptions,and contents of the NATO HV are displayed inTable 1[21−23].These clarifications provide humanrelated requirements for other EA frameworks.

Table 1 NATO HV

3.Human-centered DoDAF meta-model

The Department of Defense Architecture Framework 2.0(DoDAF 2.0) provides an overarching set of architecture concepts,guidance,best practices,and methods to enable and facilitate architecture development,which is widely adopted by governments.The data meta-model (DM2) of DoDAF 2.0 provides a mechanism to collect,organize,and store data,which stipulates the architectural data standard at a top level and thus guides different viewpoints.The data collected with the DM2 can then be rendered into either of the standard models described by the“Fit for Purpose”presentations.The complete DM2 can be found in [41].

According to the NATO architecture framework,the human viewpoint lacks specific data elements and application approaches,such as alignment achievement with human considerations.Within DoDAF 2.0,“Fit for Purpose”is“an architectural description that is consistent with specific project or mission objectives”[23].As Handley pointed out,the flexible data schema provided by the DM2 and the ability to describe custom views tailored to address specific purposes can provide the opportunity to integrate the NATO HV with DoDAF [23].

The human centered DoDAF meta-model is displayed inFig.1with its NATO HV contents.The integrity and validity of the meta-model can be guaranteed by associating the data in the HV with the concept layer data and the logic layer data in other views.The mappings between DM2 entities and HV models are listed in this figure.For example,the HV-C functions can be related to the DM2“Activity”entity;the HV-E human network can be related to the“Organization”entity.These DM2 contents can be combined with other viewpoints of DoDAF 2.0 to develop the complete EA description.Notably,although DM2 can describe the NATO human contents,the corresponding human viewpoint is lacking in DoDAF 2.0,which can be examined in future research.

Fig.1 Human-relevant DoDAF meta-model

The above human-centered meta-model can support alignment achievement in the following ways:first,from the perspective of EA planning,this meta-model determines the relationships of various viewpoints in multiple layers,which correlates business-related data and IS-related data;second,from the perspective of alignment modeling,e.g.,strategic alignment model [5],this metamodel aligns business capability,IS capability,business structures,and IS structures with the EA description;third,from the perspective of human behaviors,this metamodel tracks different kinds of performers’ relationships,tasks,and roles to capture the human uncertainty in the organizational evolution.Overall,this meta-model summarizes human-related data and provides a better foundation for business and IS alignment.

4.Human-centered dynamic alignment model

According to dynamic alignment research,the above meta-model acts as a deep structure of alignment evolution.For example,Sabherwal described the deep structure of alignment with a strategic IS management profile embracing the business strategy,the business structure,the IS strategy,and the IS structure [25],and Lyytinen explained the deep structure of IS from the perspectives of structure,actors,tasks and technology [42].This paper emphasizes human factors from a human viewpoint.Therefore,this paper considers the human-related deep structure via the DoDAF meta-model.

The dynamic alignment model is developed inFig.2,showing the first-order change and the second-order change from external and internal environments.This model explains the alignment trajectory in the organizational evolution process.These two kinds of changes occur in different complexity regions.During the organizational evolution,the organizational structure may undergo first- and second-order changes. The first-order changes may be slight,while the second-order changes are often dramatic and revolutionary.First-order adaptation and second-order adaptation can be applied in these situations.

Fig.2 Human-centered dynamic alignment model

4.1 First-order change and adaptation

According to [25] and [42],the first-order change occurs between periods of incremental adaptation,and the second-order change brings revolutionary and episodic upheaval.As Section 1 explains,the first-order change is situated in the complexity region of order,which can be adjusted by correction mechanisms,while the second-order change causes nonlinearity,which needs to be predicted,and corresponding unintended deviations should be tamed.

According to extant misalignment research [43,44],the first-order change can cause misalignment among the entities of meta-models.We enumerate the misalignment symptoms caused by first-order changes inFig.3andTable 2.For example,in regard to external changes,organizational goals and the responsibilities of business processes may be unknown (S01),which misaligns business and IS capabilities;changing business goals creates a lack of required information for performers to support their day-to-day activities (S05);and business structure and IS structure fall out of harmony with one another because business/IS actors have not participated in IS/business organizations (S10).These misalignment symptoms caused by first-order changes describe the possible deviations of the meta-model. Human factors,including goals,responsibilities,activities,skills,and groups,are mainly considered inTable 2.Overall,this kind of misalignment is easily checked and is often corrected by means of meetings,introducing or deleting elements of the model,reorganizing elements or organizations,and so on [25,45].A first-order change can be easily sensed and absorbed and always improves organizational performance after the misalignment is corrected [43].

Fig.3 Misalignment caused by first-order changes

Table 2 Misalignment descriptions

4.2 Second-order change and adaptation

The second-order change is revolutionary and involves more than one of the elements or relations shown inFig.1.For example,an IS innovation may alter systems,deploy performers,or even adjust capabilities.In the extant digitizing world,an organization’s external and internal complexities may increase with the challenge of“dancing rugged”environments [46],pervasive digital technologies [33],and a multitude of interdependent relationships[47]. Under these situations,the traditional alignment shifts away from equilibrium and starts to embrace“emergent complexity”[20].Instead of requiring a fixed solution,deviations from the intended plans generated by the second-order changes need to be“captured”and“tamed”[24].This forces governance mechanisms to identify the root causes of these dynamics and to regulate human behaviors by experimenting and learning [20].

This second kind of change can be absorbed and tamed by governance principles [20].In our previous research,we explored 10 governance principles from the extant literature [32].For example,frequent mutual communication among actors can reduce human inertia,e.g.,negative emotions due to threat perception,stickiness due to norm and value re-enactment,which can be caused by short-term success but would lead to suboptimal conditions in the long term;in addition,knowledge sharing supports the perception of unpredictable changes and is effective in organizational transformation [31];furthermore,facilitating the learning capability in a company helps to reduce uncertainties,increase companies’ internal complexity,and accelerate future change rates [48].These governance principles help to tame unintended deviations caused by second-order changes.

Overall,the whole alignment trajectory is connected by first-order changes and second-order changes.The humancentered alignment meta-model acts as a deep structure to determine alignment states.The above dynamic alignment model explains human behaviors and their longterm effects on alignment.

5.Discussion and conclusion

Human-centered architecture exploration from the alignment perspective is a novel research topic.This paper mainly introduces a human-centered alignment metamodel and develops a dynamic alignment model with this meta-model.This paper suggests further research into other human-related issues. First,the relationships between the human viewpoint and other EA viewpoints,e.g.,capability viewpoint and system viewpoint,need to be determined,and the development process also needs to be built on the basis of alignment achievement.Second,architecture data from a human viewpoint should be used to measure alignment maturity.Third,with the help of a human viewpoint,the alignment process combining topdown rational design and bottom-up emergent adaptation needs to be developed.

We contribute to the extant literature in two ways.Theoretically,we combine the EA framework and the human viewpoint to address alignment issues in the architecture design phase.In addition,we describe dynamic alignment by developing a human-centered meta-model and dynamic alignment model,which can explain first-and second-order changes and their effects on alignment evolution.Practically,the human viewpoint captures the relevant dynamics of human factors and associates them with the components defined in other views.These characteristics make it easier to analyze the human viewpoint in terms of dynamic changes in organizational transformation and evolution.This paper thus provides better support for theoretical research and practical application.

This paper still presents limitations.First,given the lack of validation of the human viewpoint,further humanrelated description models need to be developed;second,the theoretical dynamic alignment model is proposed,but the best way to guide the practical application of this model still needs to be explored.A sensing capability that can identify first-order changes and second-order changes should be considered in the dynamic alignment model.All of these limitations need to be analyzed in future research.

Overall,with the advent of emerging technologies and the associated dynamic strategies,complexity is likely to pose a long-term challenge to the analysis of alignment[49,50].The human viewpoint is a key solution to alignment complexity.Given the lack of human consideration of alignment and architecture frameworks,this paper combines traditional alignment and dynamic alignment to achieve alignment through human-centered EA design and evolution to better describe the alignment evolution process.