4.1.1 Challenges with implementing BPM
There are some challenges when implementing a BPM system in an enterprise. Riejers (2006) has identified four key factors when implementing BPM; technology, management, human and process. The BPM system has to interoperate with other systems, in order to be implemented with success in the enterprise. The BPM system should be implemented in phases and while it is also very important that the management supports the implementation. The end-users has to be ready for change, but for good implementation and usages of the BPM system, the end-users workflow have to be identify in order to setup the setup that supports the processes in the company. It is important to support the end-users, as that is a key factor for a successful implementation. The implementation team has to have a wide understanding of processes and process concepts in the enterprise and should from an early stage have a process-oriented approach to the development and implementation (Riejers, 2006). The team has to achieve an understanding of the business processes and the changing processes in the company, but it can be difficult to formalized process knowledge as it can be tacit knowledge (Al-Mashari, 2002).
4.1.2 BPM suites
BPM systems can manage processes all types of workflows. The BPM suites have different features, but the basic features of a BPM suites are; “process execution, process monitoring, customizable industry-specific templates and UI components, and out-of-box integration capabilities along with support for Web-services-based integration” (Tanrikorur, 2007, s. 1).
IT companies are offering many different BPM suites (BPMS), and each one is different solutions, solving different needs. What these systems have in common is that they are offering a way to build, execute and monitor processes (Tanrikorur, 2007).
4.2 Enterprise Resource Planning and implementation
Enterprise Resource Planning (ERP) is a software framework that is to manage a company’s throughput and integrate all data into one system (Umble & Haft, 2003). There are two major benefits for implementing an ERP system; firstly does the ERP system give a unified enterprise view of functions and departments and secondly it provides a database where business transactions are entered, recorded, processed, monitored and reported (Umble & Haft, 2003).
The ERP system works as a complex information system for the enterprise and implementing an ERP system is a very complex and high resource costly procedure and therefore does many enterprises failure with the implementation (Umble & Haft, 2003).
4.2.1 Challenges with implementing ERP
For enterprises to succeed with the implementation of an ERP system, Umble et al (2003) have identified 9 key factors for implementing ERP. Four factors are within the organizational aspect, where there has to be a clear vision and goals, commitment by the top management, excellent project management with defined a defined scope and a great implementation team. The other 5 factors are more relevant for this project and are described below.
Organizational change management
The enterprise has to be ready for changes in the structure and processes as the ERP systems are not fully compatible and implementation may require changes in key processes. Therefore the enterprise has to be ready for redesigning or realigning processes. Basically the business processes has to adjust to the ERP system if needed, and the enterprise has to be ready for these changes.
Data accuracy
There has to be control for the users entering of data, as wrong usage can lead to negative effects throughout the enterprise.
Extensive education and training
Education and training is important to ensure data accuracy, but is also important to enable users to understand the system and how it works, so they will not make workaround in the system or outside the system.
Focused performance measure Reports has to be develop that will measure the performance of how the system in performing. The management team has to set some performance measures and use it in the implementation, but also measure the performance after implementation.
Multi-site issues
When implementing an ERP system across multi-sites is two factors be in focus for successfully implementing across sites. Firstly there has to be product and process consistency across sites and secondly a centralized control over activities in the ERP system.
4.3 Merging BPM and ERP
Where ERP is an enterprise database with all business transactions, the actual workflows with a unified interface across the entire enterprise, the BPM is a way to document the processes and workflows in the enterprise (Millet, Schmitt, & Botta-Genoulaz, 2009).
Challenges with ERP systems are that they can be inflexible and lack of agility when it has to address process needs (Neelavar, 2010). The challenges with ERP are that it can be complex, difficult and resource consuming to change process in the ERP system, because all processes have to be a part of the ERP system to be performed. If there is a mismatch between the processes and the ERP setup, the enterprise will face major difficulties in their processes (Millet, Schmitt, & Botta-Genoulaz, 2009).
By implementing a BPM system the ERP system can become more agile and flexible, as the BPM have documented the processes in the company and have the track of every workflow in the enterprise (Neelavar, 2010). By focusing on formalizing the knowledge about the business processes, the BPM can ensure an overview for the enterprises processes (Sakka, Millet, & Botta-Genoulaz, 2011).
For aligning BPM and ERP there has to be an independent reference model, which can ensure the alignment (Millet, Schmitt, & Botta-Genoulaz, 2009). To achieves these advantages and ensure efficiency within the interaction between ERP and BPM it is important to have the systems aligned. In section 4.5 literatures about data processing will be reviewed in order to develop a model that can ensure the important alignment between BPM and ERP (Millet, Schmitt, & Botta-Genoulaz, 2009).
The following section will introduce the case IT architecture in the enterprise, to have an understanding of the data structure and the interaction between the two systems in DOVISTA, which is SAP as the ERP system and iGrafx as the BPM system.
4.4 Data processing
Business intelligence (BI) is a process where raw data in transformed into useful business insights, which can lead to optimization in the company and better decision-making. BI is the key technology to extract large data sets and process it to useful information. The technique for achieving BI is to process the data, where one of the techniques is data-mining (DM). The project is to process the data to develop insights, which this section will focus on (Duan, L., & Xu, L. D., 2012).
4.4.1 Knowledge Discovery in Databases
Knowledge Discovery in Databases (KDD) is about development of methods and techniques that can process data to discover knowledge in the data (Fayyad, Piatetsky-Shapiro, & Smyth, 1996). Fayyad et al (1996 s. 41) describe KDD to be “the nontrivial process of identifying valid, novel, potentially useful, and ultimately understandable patterns in data”. The importance is to find structures in the data, by getting an understanding of the data structures in each dataset. Identified patterns should be valid, novel and useful (Fayyad, Piatetsky-Shapiro, & Smyth, 1996).
According to Brachman and Anand (1996) are there nine steps in the KDD process.
1. Understand the business and data systems and set goals for the desired outcome
2. Select data sets
3. Preprocess the data
4. Identify useful datasets depended on the goals
5. Implement a data-mining method which match the goals
6. Select methods to the data-mining process
7. Perform the data-mining
8. Visualization of processed data
9. Implement knowledge in the enterprise
As identified in the KDD processes, Data mining is a part of the process and important in the KDD process. The purpose with Data mining is to extract information from data sets and transform it into usable data, which then can be used for further use in the knowledge discovery (Chakrabarti, 2006).
4.4.2 Data mining
A standard process model and methodology for Data mining projects is the CRISP-DM model, as the model describes what processes that has to be performed in order to process the data (Marbán, Mariscal, & Segovia, 2009). The CRISP-DM model defines the phases in a DM project and is separated in 6 data mining steps, where each step is defined with tasks and deliverables. Marbán et al (2009) have specified six phases in the CRISP-DM model.
First phase is to identify and understand the company’s objectives and requirements and convert the knowledge from the business into a problem definition.
Second phase is to collect the initial data and explore the data to get familiar with the data and the structures. It is also important to see if there are some data quality issues in the data sets.
Third phase is to prepare the data that has to be analyzed. The importance here is to select the data, clean it and construct the data. Here it is the purpose to construct the final datasets from the extracted raw data.
Fourth phase is the analyze phase, where the data is combined in analysis. Here is the analyze tool setup, and the data is processed in the tool.
Fifth and sixth phases are an evaluation and implementation of the results.
4.5.2 Microsoft tools for data mining
For processing data there are many tools to achieve a good data process. Business statistics have shown the immense popularity of Microsoft Excel for extracting and processing data, and develop analytical solutions. Approximately 70% of business intelligence report use Excel to generate reports (Palocsay, Markham, & Markham, 2008). When developing a reporting tool to verify hypothesis Excel is a good starting point, as it requires minimum IT support, which minimize the resources for the project (Palocsay, Markham, & Markham, 2008).
Excel provides a tool that supports data management and tools for “querying, analysis, and reporting and additional formatting, calculation and graphing” (Palocsay, Markham, & Markham, 2008).When building a report tool it should either be Excel or formatting with Excel, as adjustments is easier to do in Excel than some automate report tool (Johnson, 2010).
To clean and prepare data for analysis it should be done in a systematic way, so the processes can be done in the same way next time the data has to be processed. For systematic approach make the processing in one Excel file, but in several worksheets to have an overview over the transformation. In the process start cleaning the data, so all unnecessary data is removed and then process the data. Then copy the data into static excel file, where the setup for the analysis have been setup (Weiss & Townsend, 2005).
Microsoft Access is a database management system (DBMS) tool to structure the data and achieve a better data flexibility. The database is structured in tables, which is common to excel spreadsheets. The importance is that each row in the database has a unique ID. By creating tables in Access it is possible to join the databases in Access queries, which makes it able to multiple tables. After having made queries in Access it is possible to save it and export it to Excel and make reports with the queries. By linking the data between Access and Excel it can be accomplished that the report and queries will be updated automatic when changing the data is the Excel spreadsheets. (Chapple, 2014)
To visualize the processed data the PivotTable feature in Excel can be used. The PivotTables can be sued for organizing, grouping, and investigating data that is the intention with data (Palocsay, Markham, & Markham, 2008).