An Approach to Extract the Business Value from SOA ...

Pre-print:

N. Kryvinska, L. Auer, C. Strauss, “An Approach to Extract the Business Value from SOA Services”, Second International Conference on Exploring Services Science (IESS 1.1), 16-18 February 2011, Geneva, Switzerland, Springer, Lecture Notes in Business Information Processing (LNBIP 82), ISBN 978-3-642-21546-9, pp. 42–52.

An Approach to Extract the Business Value from SOA Services Natalia Kryvinska, Lukas Auer, and Christine Strauss Department of e-Business, School of Business, Economics and Statistics, University of Vienna, Bruenner Strasse 72, A-1210 Vienna, Austria {natalia.kryvinska,lukas.auer,christine.strauss}@univie.ac.at

Abstract. Enterprises deploy SOA for the cost-effective flexibility and enhanced usage of legacy IT. But, the shift to SOA does not come easy and cheap. Building an effective SOA platform requires tight integration between new and existing product categories, and it may require large investments. Therefore, conventional investment valuation methods need to be combined with other modern techniques to reflect SOA’s long-term strategic investment nature, the inherent uncertainty and the managerial discretion. Some emerging value-based analytical methodologies have started to be developed extensively under the umbrella concept of Business Value Analysis (BVA). However, BVA is not yet a wholly integrated and the theoretical foundation for it needs to be strengthened. Thus, in this paper, we apply modified CVP/BEP method developed under the BVA umbrella for the evaluation of SOA projects, in particular for the services reuse. Keywords: Business Value Analysis (BVA), Information Technologies (IT), Return of Investments (ROI), Service-Oriented Architecture (SOA), Break-even Point (BEP), Cost-volume-profit (CVP).

1 Introduction The enterprises are turning to SOA for the cost-effective flexibility and enhanced usage of legacy IT. But, building an effective SOA platform requires tight integration between new and existing product categories, and it may require large investments for new products such as SOA repositories. In the early stages of an SOA, limited ROI is achieved mainly from the creation of services from existing applications. It is not unusual for some organizations to experience negative ROI compared to conventional software development methods because additional investment may be required to achieve reusable services and to build an SOA infrastructure. IT flexibility becomes a reality once there are enough services in an SOA to stimulate extensive internal and external services reuse and consumption (Fig. 1). But, the largest business value of SOA, e.g., business agility and competitive advantage, is only gathered once the business starts to use services to orchestrate processes within the enterprise and choreograph processes between enterprises. Unlike traditional technology investments that tend to be implemented enterprise-wide in “big-bang” fashion, SOA is realized

over a long time horizon through many distinct projects distributed throughout an organization (Fig. 2). And, in addition to the uncertainty associated with future service reuse, quantification of the soft benefits of SOA is often problematic [1 ÷ 4].

Number of Services/Projects

Fig. 1. Benefits from Components Reuse in SOA [5]

Therefore, the conventional valuation methods need to be combined with concepts modern techniques to reflect SOA’s long-term strategic investment nature and the inherent uncertainty involved in SOA investment. These and some other emerging value-based analytical techniques, showing a promise, have started to be developed extensively under the umbrella concept of Business Value Analysis (BVA). BVA attempts to analyze the factors and forces that will shape the future instead of trying to forecast the future. It brings together methodologies that extend DCF (Discounted Cash Flow) and other traditional financial analysis techniques to include intangibles and other factors common to the digital economy. BVA includes the following techniques: Real-Options, Business Model Dynamics, Synthetic Markets, Multicriteria analysis, Ratio methods, Break-even Point and Portfolio theory [6 ÷ 8].

Fig. 2. Increase of SOA “Business Value” with Number of Projects [9]

However, BVA is not yet a wholly integrated analysis technique and the theoretical foundation for it needs to be strengthened. Further, more work needs to be done on

the practical side. Properly implemented with supporting IT/SOA, BVA has the potential to lead to better investment decisions and business results (Fig. 3) [6, 10, 11]. Thus, in this paper, we apply modified CVP/BEP method, developed under the BVA umbrella, for the evaluation of SOA projects; in particular for the services reuse. Reduced Operational Risk

Reduced processing time Reduced errors

Improved Ability to Change Enable new product development Improved flexibility

Reduced system downtime

Flexibility value drivers Profitability value drivers Protect existing revenue Generate new revenue Increase existing revenue

Ease of integration

Reduced time-to-market

Increased reuse

Reduced integration time

Reduced integration cost

Increase Revenue

Increased Profitability

Decrease Costs

Reduced maintenance cost

Improved ability for compliance

Fig. 3. SOA “Business Value” Features Tree [12]

2 SOA Concept and Its Importance for Enterprise Management The Service-Oriented Architecture (SOA) is an architectural framework that takes business applications and breaks them down into services that can be made available for use independent of the applications and the computing platforms on which they run. These services can be integrated and used to build new capabilities supporting new functionality from within current portfolio or from extended value chain [12, 13]. Today’s businesses depend on electronic processes at every level. An organization’s ability to stay competitive relies heavily on being able to adapt its electronic processes in support of initiatives designed to improve productivity, reduce costs, deliver higherquality information, and accelerate routine tasks. However, adapting business-critical processes requires evolving the systems they run on quickly and cost-effectively. This, in turn, is an extremely multifaceted task [4, 14, 15]. For these enterprise needs, SOA offers protocol independence, meaning that different consumers of computing services - such as an application, a server system, or a human end user can communicate with the same service in a different way to obtain the data or functionality desired. Services themselves function as discrete components, designed to aggregate underlying complex computing interactions into reusable “packages” that can be invoked whenever that particular piece of functionality is required (Fig. 3) [16, 17].

3 SOA Investments and Business Value Estimation Methodologies Any methodology has to be designed to assess the business “competitive value assets” that are important to effectively implement a business strategies; assets such as those used to understand customer needs and turn them into knowledge, those used to create value, and those needed to deliver value. And, since an assessment process is in itself a strategic initiative, it has to be designed and implemented carefully to fit the goal that each benefit can provide. It may be possible to use the same evaluation to gain multiple benefits but it must be designed to do so [18]. As enterprises grow, it typically takes longer to see the impact of actions or investments. The effect also may be distributed throughout the enterprise, requiring more sophisticated techniques to capture the organizational impact of a particular action or investment. The distribution of effects may also make it more difficult to capture the right measures or collect data on measures in a timely or consistent way, since the effects are likely occurring outside the IT unit’s domain. Thus, from an organizational perspective, it is difficult to place a concrete value on many decisions, actions, or investments because the effects are too distributed or too moderated by other factors to measure them easily or accurately [19 ÷ 23]. 3.1 Problems with Measuring SOA Related Value What is it about SOA that makes measuring its strategic value so difficult? A simple answer is that there are too many ways to measure it. Every consultant, practitioner, or faculty member has a unique method for capturing SOA value. Table 1 provides the list of sample measures [19]. Many of these models were designed for the private sector rather than for large enterprises, and some are more easily applicable than others to the goals and objectives found in the academic environment. Even if IT leaders can sift through these models and find an appropriate one to use, the unique challenges of IT valuation have only begun. Table 1. List of Methods for Valuing IT in general and SOA in particular [20] Accounting rate of return

Boundary value

Multi-objective methods

Analytic hierarchy process

Costs-benefits analysis

Profitability index

Application benchmarking

Costs-effectiveness analysis

Process quality management

Balanced scorecard

Costs-value techniques

Schumann’s method

Bayesian analysis

Costs-revenue analysis

Real-options valuation

Bedell’s method

Delphi evidence

System dynamics analysis

Buss’s method

Information economics

Utility function assessment

Benefits-risks portfolio

Infrastructure analysis

Ward’s portfolio analysis

Break-even analysis

Investments portfolio

Wissema’s method

And, why is measuring the value of SOA so hard? As SOA has permeated organizations, measuring its direct contribution to the bottom line has become more

difficult. This is because cost savings, primarily in the form of headcount reduction and productivity improvements, were gains that were achieved early on in a system’s lifecycle. In fact, as SOA is evolving over time, the difficulty of calculating the value of SOA investments is increasing, due in large part to refocusing application development away from clerical efficiency and toward a wider range of less well-defined outcomes such as competitive advantage, knowledge management, and improved organizational performance [20 ÷ 23]. 3.2 Splitting the Variables of SOA “Business Value” into the Dimensions We distinct and split here the SOA value variables into two dimensions: costs and benefits with 6 cost variables and 5 benefit variables. This splitting is needed for the definition of hypothesized relationships between the costs, benefits, and the “Business Value” (Fig. 4). The costs dimension is related to the efforts required for an enterprise-wide IT effort such as SOA, while the benefits are potential rewards of the SOA deployment. Costs Organizational

Business Process

Integration

Technology

Servicification

Human Recources

SOA Business Value

Benefits Services/ Recources Re-use

Human Recources

Architecture Flexibility

Integration

Open Standards

Fig. 4. Costs and Benefits Components of SOA “Business Value” [24]

The most important benefits are derived from low vendor lock-in, allowing enterprises to replace customized applications and products. This is followed by variables such as platform and technology independence leading to greater collaboration and integration, with less spending on middleware infrastructure, time to market, which enables enterprises to come out with products/services in a shorter timeframe, adaptability to emerging business scenarios. The benefits due to a reduction in application redundancy are emerging from rationalization of applications, and re-use of business processes, applications, and infrastructure through the potential reuse of services and infrastructure, standardized employee skill sets. As the SOA interfaces are abstracted from the implementation (loose coupling) it enables systems to be assembled and disassembled easily as they have less dependency on other

systems. Finally, there are benefits from reduced cost and time of introducing new business processes, reduced cost and time of modifying existing business processes, reduced cost and time of introducing new applications, reduced cost and time of modifying existing applications, reduced cost and time of introducing new IT infrastructure, and reduced cost and time of modifying existing IT infrastructure. To conclude, the IT investment decision makers need to have a comprehensive understanding of the cost and benefit variables to arrive at the business value measure to justify SOA-based IT architectures [20, 22, 24]. 3.3 Risk Management Challenges within of IT/SOA Related Projects Enterprises are investing capital into certain objects in order to generate cash inflows and subsequently to increase the return of the invested capital. Typically risk-averse management is making risky investments hoping to achieve an excess return over the risk-free rate. There is a general connection between risk and return of an investment object: higher return is systematically associated with higher risk [25]. Going further into the IT investments direction, we need to understand the context of these projects and the risks associated with them, since IT related investments are ones of the most profitable as well as the most challenging for modern enterprises. Thus, any IT investment project has to be embedded in an organization’s technology infrastructure, relevant business processes, organizational environment, and external relationships. As an example, the risk factors that can impact investment decision process: • Organizational factors - misalignment of internal goals, lack of leadership support; • Business process factors - fear of changing work assignments; • Technology factors - rapid changes in technology, interacting with parallel systems, scale and complexity [7]. Furthermore, project failures are the result of the multiplicity of risks inherent in software project environment. Software development projects are collections of larger programs with many interactions and dependencies. It involves a creation of something that has never been done before although the development processes are similar among other projects. As a result, these projects have a dismal track-record of cost and schedule overruns, and quality/usability problems [26, 27].

4 BEP/CVP Analysis with Emphasis on Services Re-use in SOA Cost-volume-profit (CVP) analysis expands the use of information provided by breakeven analysis. A critical part of CVP analysis is the point where total revenues equal total costs (both fixed and variable costs). At this breakeven point (BEP), a company will experience no income or loss. This BEP can be an initial examination that precedes more detailed CVP analyses. The conventional Cost-volume-profit analysis (Fig. 5) employs the same basic assumptions as in breakeven analysis. Further, the breakeven point can be determined with a mathematical equation, using contribution margin, or from a CVP graph [28, 29].

Total Revenue/Costs

Total revenues line

Breakeven Point

Total costs line Operating income

Operating loss Number of Services

Fig. 5. Conventional CVP/BEP Approach [29]

4.1 Short Overview of Reference Resources Since its development by Jaedicke and Robichek [30], the various CVP models have been proposed and studied: single-product versus multi-product, single production technology versus multiple technologies, single uncertainty source versus multiple uncertainty sources, uncertainty with respect to price versus uncertainty with respect to sales quantity, the assumption that production equals sales versus differentiation of the production quantity from the sales quantity, specification of the decision question simply as produce-not produce versus determination of a quantity to produce and/or a price to set, use of the fundamental CVP equation alone versus the addition of an “economic” demand function relating quantity sold to price charged, and so on and so forth [31]. Namely, the following researchers have contributed to the cost-volume-profit analysis under uncertainty: Dickinson [32]; Hilliard and Leitch [33]; Kottas and Lau [34]; Ismail and Lauderbeck [35]; Constantinides, Ijiri and Leitch [36]; Barry, VelezArocho and Welch [37]; Lau and Lau [38]; Kim, Abdolmohammadi and Klein [39]; and Yunker [31, 42]. 4.2 Applying Modified CVP Approach to SOA Service Reuse Projects The purpose of our work was to apply modified CVP/BEP method for the evaluation of SOA projects in general and for the services reuse in particular. We take as basis a diagram presented by [40] (Fig. 6), which reflects the specific of SOA investments in comparison with traditional IT software development approaches. From the Fig. 6, we can see that the SOA-related service costs do not have linear nature, unlike IT traditional service implementation costs. This happens because the services developed within of SOA framework can be re-used further. Consequently, the number of (business, composite) services grows but the costs for their development – not or almost not, since they are just orchestrated/reconfigured from already existing ones. Thus, according to what was discussed in the introduction, in order to find BreakEven Point (e.g., the point where SOA-based approach to the services development

Service Development Costs

and deployment become economically feasible in comparison to the conventional software development methods), we perform here an analysis of BEP positioning depended on the following aspects: amount of services re-used, size of SOA projects and enterprises, specific implementation areas and industries. Traditional Approach

Economic Advantages

SOA Fixed Cost

SOA Approach New business capabilities become economically feasible

Break-even Point

Number of Services

Fig. 6. Economic Value of SOA-oriented Development [40]

Besides, we apply combined/hybrid approach. Namely, for the conventional software/services development projects we apply conventional CVP/BEP analysis (Eq. 1) widely described in the mentioned above references as well as in [41] and [42]. But, for the SOA-based projects, we use modified cumulative distribution function (cdf) (Eq. 3 and 4) of the negative exponential distribution (Eq. 2). Why do we apply the modified cdf for the SOA costs? From the Fig. 6, we can see that SOA first project costs (e.g., SOA fixed costs) are not equal to zero. In comparison to the conventional IT “single application” approach – to build SOA framework considerable investments are needed, see details in the introduction. So, we set up the value “A” - SOA fixed costs, which is higher than “1”.

π = pq − cq − f = ( p − c )q − f

(1)

where: .. .. profits, p .. price, q .. quantity, c . . unit cost (average variable cost), f .. fixed costs [31, 41, 42].

⎧λe − λx , x ≥ 0 f (x ) = ⎨ x 0

⎧1 − e − λx , F (x ) = ⎨ ⎩ 0,

x≥0

⎧ A − e − λx , F (x ) = ⎨ ⎩ 0,

x≥0 x