University of Economics, Prague
Faculty of Informatics and Statistics
DECISION-MAKING PROCESS ON THE IMPLEMENTATION OF DATA ANALYTICS SOLUTIONS IN THE SMALL AND MEDIUM-
SIZED BUSINESS
MASTER THESIS
Study programme: Applied Informatics Field of study: Information Systems Management
Author: Volha Bahlai
Supervisor: Ing. Martin Potančok, Ph.D.
Prague, June 2021
Declaration
I hereby declare that I am a sole author of the thesis entitled “Decision-making process on the implementation of data analytics solutions in the small and medium-sized business”. I duly identified all citations. The used literature and sources are stated in the attached list of references.
Prague (Date)... Signature Volha Bahlai
Acknowledgement
I hereby wish to express my appreciation and gratitude to the supervisor of my thesis, Ing.
Martin Potančok, Ph.D. for expert help with the topic and constructive comments.
I also want to thank FIS Master Program Coordinator FIS Master Program Coordinator Mgr. Veronika Brunerováa for the constant support, clarifications and addressing emerging issues.
I would also like to thank the International Visegrad Fund for the opportunity to study the Information Systems Management program.
In addition, I would like to thank the employees of the companies, where the case studies were conducted, for their time and operant response.
Abstract
The opportunities and benefits that Data analytics provides increase the interest in its implementation by small and medium-sized businesses. The use of different solutions allows improving companies’ performance and is a strong competitive advantage. However, the problems and challenges the SMBs face with data analytics implementation significantly decrease the number of successfully completed projects.
In the research, the aim is to design a decision-making model for data analytics implementation and development for SMBs. Based on theoretical and practical findings for the particular companies, the model provides guidance on making decisions on the project initiation stage and reduce the likelihood of the project failure.
The designed model is used to analyse five data analytics projects for implementation for two companies. As a result of using the model, the companies were presented with evaluating the degree of readiness for implementation, the alternatives of project implementation solutions and their evaluation plan. Due to the provided findings, the companies are able to reveal that some projects cannot be allowed to be implemented yet or as planned. It is also discovered that the general technical preparation of the companies is important for implementing data analytics projects. Despite the positive results of model validation in companies, the usage of the model by other companies needs future verification due to the different maturity levels and companies industries, level of analytics uses and other factors.
Keywords
Data analytics, Implementation process, Decision-making process, Small and medium- sized business, Decision-making model, Data analytics solutions
JEL Classification
O32: Management of Technological Innovation and R&D M15: IT Management
M00: General (M: Business Administration and Business Economics • Marketing • Accounting • Personnel Economics)
Content
1 Introduction ... 8
1.1 Data analytics definition ... 8
1.2 Specification of SMBs for Data analytics usage ... 10
1.3 Data analytics implementation benefits and challenges ... 11
1.4 Research description ... 15
1.4.1 Aim and objectives ... 15
1.4.2 Thesis Structure ...16
1.4.3 Research Limitations ...16
1.4.4 Methods description ...16
2 Data analytics implementation literature review ... 20
3 Company's Data analytics maturity assessment ... 24
4 Business understanding and project formulation ... 27
5 Analysis of Data analytics structure elements ... 30
5.1 State-of-the-art tools and techniques ... 30
5.2 Data-savvy people ... 31
5.3 Processes ... 32
5.4 Data ... 33
5.5 Opportunity to work with Big Data ... 35
6 Data analytics ecosystem ... 37
7 Data analytics solutions for SMBs ... 43
8 Costs and benefits measurement ... 49
9 Decision-making model ... 51
10 Final solutions ... 55
10.1 Projects evaluation for the Case 1 ... 55
10.1.1 Project 1 ... 55
10.1.2 Project 2 ... 57
10.1.3 Resume ... 58
10.2 Projects evaluation for the Case 2 ... 58
10.2.1 Project 1 ... 58
10.2.2 Project 2 ... 60
10.2.3 Project 3 ...61
10.2.4 Resume ... 62
4
Conclusion ... 63
List of references ... 64
Annexes ... 70
Annex A: Maturity level analysis ... 70
5
List of Figures
Figure 1. Depicting relations of fields, an Euler diagram (Valchanov, 2018) ... 9
Figure 2. Organisation structure of the company, Case 1 (Author, 2020) ... 18
Figure 3. Organisation structure of the company, Case 2 (Author, 2020)...19
Figure 4. AIMS-BI methodology (Rao-Graham, McNaughton and Mansingh, 2019)...21
Figure 5. Business and data ecosystem transformation (Jackson and Carruthers, 2019) .. 30
Figure 6. The data value chain and life cycle (Bianchini and Michalkova, 2019) ... 32
Figure 7. Key dimensions and data types (‘The Data Value Chain’, 2018) ... 33
Figure 8. Information supply chain (Wells and May, 2017) ... 37
Figure 9. Traditional data analytics architecture (Kayay, 2020) ... 37
Figure 10. Modern Analytics Ecosystem (Wells and May, 2017) ... 40
Figure 11. Data analytics architecture of the Case 2 (Author, 2020) ...41
Figure 12. Data analytics architecture of the company Case 1 (Author, 2020) ...41
Figure 13. Maturity evolvement (Elliott, 2018) ... 43
Figure 14. Magic Quadrant for Analytics and Business Intelligence Platforms (Richardson et al., 2020) ... 46
Figure 15. Project costs compared with benefits (Laursen and Thorlund, 2016) ... 50
Figure 16. Decision-making model (Author, 2020) ... 51
6
List of Figures
Tab. 1 Data description for Case 1 (Author,2020) ... 34
Tab. 2 Data description for Case 2 (Author,2020) ... 35
Tab. 3 Summary of Differences Between Traditional and Modern Platforms (Parenteau et al., 2015) ... 38
Tab. 4 Model outputs, Project 1 Case 1 (Author,2020) ... 56
Tab. 5 Model outputs, Project 2 Case 1 (Author,2020) ... 58
Tab. 6 Model outputs, Project 1 Case 2 (Author,2020) ... 59
Tab. 7 Model outputs, Project 2 Case 2 (Author,2020) ...61
Tab. 8 Model outputs, Project 3 Case 2 (Author,2020) ... 62
7
List of abbreviations
BI Business Intelligence
SMBs Small and Medium Sized Businesses NLQ Natural Language Generation NLG Natural-language generation AI Artificial Intelligence
KPI Key performance indicators SaaS Software as a service
CIS Commonwealth of Independent States
8
1 Introduction
Digitalisation penetrates all spheres of public life, including business, and generates vast amounts of data that can create significant value. However, data in itself is meaningless without using the analytics sphere’s developments (Sedkaoui, 2018). Benefits from using data can be realised after the data is transformed to the information that can be analysed and drive decision-making and strategic actions (Chu, 2003; Sherman, 2014). Digitalisation increases the speed and dynamics of appearing and processing this data, information and knowledge creation (Rao-Graham, McNaughton and Mansingh, 2019).
The more ways of using the information for different purposes appear in the company, the more value the same information can create (Loshin, 2012). Consequently, understanding the benefits of analytics increases spending on data analytics implementation and development.
The survey of 1518 executives in 2016 (EY Data Analytics Report, 2017) revealed that more than half of the companies are going to invest at least 10 million dollars in data and advanced analytics resources in 2 years.
Using data analytics leads to active data analytics solutions development, the emergence of new technologies in this market, and simplification of its implementation process. Previously, the implementation of classical warehouse and analytical systems was financially and technically complicated, and SMBs were often challenged by the complexity and the resources the implementation required (Rao-Graham, McNaughton and Mansingh, 2019). The current trend is to provide end-to-end workflows by blurring the distinction between the data and analytics solution markets (Goasduff, 2020). Consequently, the availability of such technologies as cloud computing or access to tools using SaaS technologies simplifies the implementation and development process, reduces efforts it requires and opens up opportunities for using advanced data analytics by SMBs.
The interest in data analysis and analytics solutions increases during crisis periods. According to the survey results provided by Sisense Inc. (State of BI & Analytics Report, 2020), during the COVID-19 worldwide epidemic, 34 per cent of companies are going to increase their investments in data analytics tools. The SMBs lead this tendency by using their advantages in mode flexibility to be competitive with large companies.
1.1 Data analytics definition
Data Analytics represents a complex concept that includes activities from different science fields. However, there is a lack of generally accepted definitions, and different interpretations of key concepts in this field mislead the general understanding of the phenomenon. Business Intelligence, Data Analytics, Business Analytics are ‘often used interchangeably in the academic and trade literature as well as by practitioners’ (Rao-Graham, McNaughton and Mansingh, 2019). This implies the need for a careful selection of the appropriate definitions under this research to ensure correct sources findings.
9 One of the basic definitions is that data analytics uses ‘statistical and mathematical data analysis that predicts what scenarios are most likely to happen’ (Sedkaoui, 2018). Even so, these two concepts are not only the fields used in analytics. Sheikh (Sheikh, 2013), in his definition, also includes data mining and mentioned visualisation as crucial parts of the concept for creating the vision of future possibilities. At the same time, he clarified that the visualisation process is a reporting mechanism that helps make the decision more acceptable for understanding but not for decision creation.
Both definitions also declare that data analytics results provide information about the future and the possible ways of action based on the available data. Nevertheless, the understanding of data analytics as analytical techniques and their implementation for future understanding and prediction is rather narrow. The broader definition describes data analytics as an overarching science or discipline that encompasses the complete management of data (Brown, 2019). It means that it contains a process of analysis and providing results and data collection, organisation, storage, and all the tools and techniques used for this purpose. The broad meaning of Data Analytics will be used within the framework of this work. It means that it includes a range of ‘approaches and solutions, from looking backwards to evaluate what happened in the past to looking forward to doing scenario planning and predictive modelling’
(Data Analytics, no date).
However, such broad explanations create the problem of differentiating concepts such as Data Analytics, Business Intelligence, Data Science and Business Analytics. Figure 1, an Euler diagram, shows how these terms are interconnected and have a lot of common features.
Figure 1. Depicting relations of fields, an Euler diagram (Valchanov, 2018)
According to the diagram, Business Intelligence can be described as a part of Data Analytics.
Negash and Gray (2008)(Rao-Graham, McNaughton and Mansingh, 2019), provide the definition of Business Intelligence that is very close to Data Analytics. BI is ‘combining data gathering, data storage, and knowledge management with analytical tools to present complex and competitive information to planners and decision-makers’. Another definition, provided by Sedkaoui (Sedkaoui, 2018), describes business intelligence as a ‘combination of
10 applications, infrastructure, tools and best practices that enable access to analysis of information to improve and optimise decisions and performance’. According to these two definitions, such characteristics as data management, using data analysis, support of the decision-making process is typical for both Data Analytics and Business Intelligence. However, Sheikh (Sheikh, 2013) pays attention to the fact that Business intelligence does not provide information about the future and answers the questions about the past and describes it as a reporting and data warehousing activity.
Data science is a field that focuses on the technical aspects of data analysis. It is defined as ‘a set of principles, problem definitions, algorithms, and processes to extract non-obvious and useful patterns from a large data set’ (Kelleher and Tierney, 2018). Such activity is focusing on finding new by using a dataset rather than answer specific questions.
The term Business Analytics is often used as an alternative to Data Analytics. Gartner (Gartner, no date) defines Business Analytics as a combination of ‘solutions used to build analysis models and simulations to create scenarios, understand realities and predict future states’.
They specify that it includes data mining, predictive and applied analytics which are also a part of Data Analytics. However, Business analytics is more focused on applying business logic and processes around business data ‘to draw information that can be used by different levels in the organisation’ (Bag, 2016). Thus, despite its closeness to the concept of Data Analytics, Business Analytics is more business-focused and includes a broader range of business logic examination and analysis than mathematical and statistical models.
1.2 Specification of SMBs for Data analytics usage
Current research focus on the Data analytics for companies that defines as a small and medium businesses. The Small and medium-sized businesses are the largest group of companies in terms of the number that two main features can characterise: size and resources (User guide to the SME Definition, 2020).
Size: It includes the number of employees, turnover value and balance sheet total. It is the most commonly used attribute of the companies used in different countries to define the company. Small companies describe with a number of employees up to 250, the turnover for such companies is under 50 million euro and annual balance sheet total not exceeding 43 million euro.
According to this feature, the companies can also be divided into three categories (User guide to the SME Definition, 2020):
Micro-enterprises. The companies employ fewer than ten persons and whose annual turnover or annual balance sheet total does not exceed EUR 2 million.
Small enterprises have less than 50 employees, and annual turnover does not exceed EUR 10 million.
Medium-sized enterprises. It is a company with a number of employees less than 250, the annual turnover is not exceeding EUR 50 million, or an annual balance sheet not exceeding EUR 43 million.
11 Resources: Access to significant resources can influence companies’ performance and competitiveness and be not defined as SMB for this purpose. Resources include ownership, partnership and linkages. The company should be financially autonomous, not link to other companies, and be totally independent if the holding earns not more than 25% of shares or voting rights.
These features specify the Data Analytics projects implementation and require different managerial approaches in comparison with other types of companies (Rao-Graham, McNaughton and Mansingh, 2019). SMBs operate with limitations discussed above and in a very competitive environment. First of all, they compete with big companies that have all the resources they need to manage with high-level technical solutions. Nevertheless, SMBs can successfully compete with a big one, and one of their strengths is agility (Rao-Graham, McNaughton and Mansingh, 2019). It means companies can quickly react and adapt to the new conditions on the market to provide a demand product or service. At this point, large companies struggle with negotiation and deploying decisions after analysis and formulation (Lawson, Hatch and Desroches, 2019). The usage of advanced analytical solutions provides SMBs with the ability to combine their agility and flexibility and create a competitive advantage while large companies struggle with formal structure and bureaucracy (Rao-Graham, McNaughton and Mansingh, 2019).
On the other hand, such companies compete with startups that are even more agile and start to attract markets from the start (Lawson, Hatch and Desroches, 2019). Nevertheless, technology can also become a competitive advantage for SMBs in this case. It is often hardly possible for startups to justify investing in technologies and qualified people for their operation (Small businesses, big technologies. How the cloud enables rapid growth in SMBs, 2014).
They need to wait before they start to create stable revenue to ensure investors that such investments are reliable. Here, SMBs have much more financial and labour resources to implement such projects. The exception is startups, for which business processes data analysis is the critical component for product or service delivery.
The appearance of new technologies that SMBs can use contribute to significant digital transformation of such companies. According to SMB companies’ study, the significant positive correlation was found between data analysis and reporting and its productivity, profitability measures such as the EBITDA and ROE (Bianchini and Michalkova, 2019). One of the sides of the transformation is ‘responsible data management to fully exploit their agility’ (Supporting specialised skills development: Big Data, Internet of Things and Cybersecurity for SMEs, 2019) and create a data-driven company. This domain has an impact on using Big data, cybersecurity management and digital reality, such as Artificial Intelligence.
The development of cloud technologies is another impetus for SMBs technological changes. It significantly simplified access to the IT infrastructure, reduced costs and minimised requirements for its maintenance. The result is easy access to the information and its usage.
1.3 Data analytics implementation benefits and challenges
The high interest in Data analytics implementation conditioned by the long-lasting advantages it provides for any kind of business. The implementation of Data analytics requires the usage of systems that significantly change speed and efficiency compared with manual work, leverage
12 the time and geographical limits and provision of the results in real-time according to the events that are clearly defined (Sheikh, 2013). In the report about using data analytics, KMPG (Boshuizen and Elder, 2016) highlights the following key outcomes that data analytics provides:
The ability to process a large amount of data efficiently and consistently;
Finding inconsistency between business process and transactional data;
Using predictive analytics for decision-making based on historical trends.
These outputs could have an impact on five channels inside the company (Bianchini and Michalkova, 2019) : further research and development with data driven decisions, developing new products and services where the data can be or product itself or the main resource, optimisation of business processes and making them data-driven, improving or developing new management approaches. According to the KPMG survey (Coops et al., 2015), companies mostly use data analytics for risk management, sales and marketing, financial management and research and development (R&D) processes. The most visible benefits companies receive from using Data Analytics for marketing purposes is by improving dialogues with customers and predicting future trends. Simultaneously, the most growing segments are about using analytics in human resources, financial statement audit and capital allocation.
The benefits can also be divided into current and future regarding the value they created. When the current provides results in a short term perspective, future benefits create value that can be evaluated after a long time. According to the survey (Going beyond the data: Achieving actionable insights with data and analytics, 2014) the current benefits are related to fast and accurate decision-making, improving dialogue with customers, increased employee satisfaction, actionable insights, better targeting of the products and reduction of business risks. The future potential benefits are individualised marketing activities, identification of the new revenue streams and prediction of the future market trends.
Each project, including the data analytics solution implementation, launches to solve some particular problem or need the company faces. Therefore, the company expects that the project will bring value and meet the goal. For this reason, companies make precise calculations and informed decisions at the first step of initialisation before any project starts. However, a significant number of different data analytics projects are considered failed after evaluating their effectiveness after their implementation. For example, Slánský (Slánský, 2018a) indicates that about 50% of standardised BI reports, created after data analytics implementation, are not used at all. Capgemini Consulting found out that only 27% of executives described their Big Data initiatives as “successful” (Colas et al., 2014).
High risks and challenges associated with data analytics implementation jeopardise the investment’s effectiveness, especially for SMBs. The low level of ICT understanding and lack of skills, high costs of acquiring analytics technologies, low quality of data are some of the issues SMBs companies meet. The Standish Group CHAOS report (Ahmed and Pathan, 2018) provides that only 28% of projects in SMBs are completed on time and with defined budgets, and 31% are cancelled.
To reduce and prevent the number of failures, a large number of different factors should be considered in the initial stage of the project, especially for SMBs. Rao-Graham, McNaughton and Mansingh (Rao-Graham, McNaughton and Mansingh, 2019) highlight five problems that
13 distinguish SMBs in the process of analytics implementation that also proved by other researches:
Low level of ICT literacy and awareness by the management. It can be a lack of understanding of the needs for changes in the traditional business because of little knowledge of benefits and lack of trust compared to costs (Bianchini and Michalkova, 2019; Supporting specialised skills development: Big Data, Internet of Things and Cybersecurity for SMEs, 2019). Besides, the size of the company influence the visibility of the added value due to the small turnover (Supporting specialised skills development: Big Data, Internet of Things and Cybersecurity for SMEs, 2019).
Issues with business questions.
The costs of acquiring technology and luck of suitable financial options (Bianchini and Michalkova, 2019). According to a Standish Group CHAOS report (Ahmed and Pathan, 2018), the over-budget for such projects reaches 189%. Because companies do not have resources for experiments, the solutions have to serve the problems on hand (Supporting specialised skills development: Big Data, Internet of Things and Cybersecurity for SMEs, 2019).
Luck of analytical skills and human resources. Such companies are limited in attracting and retaining the high demanded specialists needed for data analytics implementation (Bianchini and Michalkova, 2019; Supporting specialised skills development: Big Data, Internet of Things and Cybersecurity for SMEs, 2019) and the creation of centralised planning and centres of analytical excellence (Lawson, Hatch and Desroches, 2019).
Data quality and volume issues. Their ability to make strategic decisions using analytics struggle with the time SMBs need to collect enough data for such decisions (Lawson, Hatch and Desroches, 2019) and their ability to collect and process it (Bianchini and Michalkova, 2019).
The problems SMBs face in data analytics introduction are similar to the adoption of other new technologies as a part of digital transformation. Organisation for Economic Co-operation and Development (Bianchini and Michalkova, 2019) also adds to the already mentioned list some external barriers for SMBs. These barriers are:
Access to investments and external financing.
Availability and costs of external data that can give more value to the existing internal.
The complexity of data protection regulation in the domain of personal data and high compliance risks.
Luck of solutions destined for SMBs. The existence of the new ICT products does not fully take the specific needs of SMBs into account in terms of price and functionality.
Companies are also struggling with the integration and adoption of the products into existing infrastructure (Chandler et al., 2011). It is necessary to ensure that platforms for implementation fit the context of the current architecture, processes and people because such a solution will be more acceptable for end users.
The barriers in Data analytics implementation causes a lot of problems for the company and disappointment in the Data analytics projects. Slansky (Slánský, 2018a) defines a big number of consequences that can appear because of the appropriate integration of the data analytics
14 solutions. Here is the summary of them that can be taken into account from the beginning to be avoided:
Parallel elements of Data Analytics infrastructure. Problems that companies face with Data Analytics usage, like unclear outcomes, difficulties with usage, etc., can force different departments to use their own Data Analytics environment. It leads to the multiplication of both data and architecture in one organisation and increases tools support costs.
Parallel/local organisations with the Data analytics environment.
Duplication of the teams causes the same problem as with the data architecture. The number of people assigned for such projects can be much higher than it needs and leads to loss of control under the content.
Multiplication of content. Duplicating the environment and teams leads to creating too much content with different results due to the use of various methods, data and sources. As a result, users can not understand the origin of the output and which of the output they can trust.
Ignorance of the parts of the environment. The lack of information about the current environment and its functionality challenges the new developments. Some of the components can be not visible and should be described and stored for future analysis. Lack of information about existing components influences the requirements and general idea of the solution for implementation. As a result, it creates unrealistic expectations. Another issue of the current structure ignorance and insufficient business analysis is the creation of unnecessary content on which companies waste resources but which are not used.
Insufficient requirements. Precise requirements based on the detailed analysis of the business help avoid a significant number of problems and reduce dissatisfaction with the result. However, requirements can be defined incorrectly, or they can be changed while the project is already running. As a result, the project can vary significantly and do not provide expected results.
The inflexibility of future solution development and restrictions in managing solutions in a unified manner. The project’s ownership and governance should be clearly defined to ensure that it is not led toward achieving the goals. Failure to comply with this rule can lead to a possible situation when the architecture does not allow quick changes because it can be not controlled enough to provide such changes. Another cause is the inflexibility of the architecture and, as a result, the lack of possibility to extend it.
Growing distrust of the outputs of other departments and the ability to provide benefits. The disappointment and distrust of data analytics often connected with communication on different levels. First of all, the expectations could be miscommunicated from the beginning with the possibility of subjective perceptions.
Another number of causes relates to the insufficient description, explanation and visualisation of the project to the end-users. The lack of clarity in the Data Analytics processes creates many questions and mistrust, primarily when the output is based on advanced algorithms.
15 Increase of time, costs, resources and the efforts required. All the problems companies face during Data analytics implementation, as a result, affects the time, costs, resources and efforts.
However, some objective limitations should be taken into account when planning an implementation project (Sheikh, 2013):
Dependence of the project from the quality of input data. Variables can become irrelevant during the time because of changing business conditions. As a result, models cannot be reliable because automated decision-making systems can not know about it.
The scenarios of automated decisions should occur according to real-time events. Slight differences in events may affect strategies and solutions proposed by systems.
The existence of these limits shows that any Data analytics system should be controlled and adjusted each time the changes in real-life appear. However, it is difficult to predict all the real-world situations to ensure that provided results and decisions will occur as intended.
To lower the likelihood of the negative impact, Brown and Lockett (Rao-Graham, McNaughton and Mansingh, 2019), in their empirical investigation, found the crucial role of the intermediaries in the adoption process and also a preference of the aggregated multifunctional applications for implementation. Intermediates can provide services of financial partnership, knowledge dissemination, technical advice and solution provision. An example of the application can be a SaaS solution that provides ICT service at a lower cost and without the necessity to manage it for themselves (Rao-Graham, McNaughton and Mansingh, 2019).
1.4 Research description
1.4.1 Aim and objectives
The indicated specifics, challenges and limitations should be taken into account at the early stages of Data Analytics development to reduce uncertainty and the project's risk of failure.
Thus, SMBs need to find proper tools and methods to evaluate and analyse planned data analytics implementation projects that take into account the factors affecting the success of such projects.
The aim of the research is to design a decision-making model for data analytics implementation and development for the SMBs to provide guidance on making decisions on the stage of the project initiation and analysis.
To achieve the aim of the research, the next objectives should be delivered:
1. Analyse the theoretical basis of the data analytics implementation process in the enterprises with a focus on SBMs.
2. Analyse the current market of data analytics solutions and their validity for SMBs.
3. Develop the decision-making model for data analytics implementation for SMBs.
4. Provide the results of analysis for the data analytics implementation for specific companies within a case study using the developed model.
16 The work is supposed to contribute to the development of the topic of implementation of data analytics projects. By the insights and findings, the researcher aim to add to the topic for future improvement and simplification of methods for managing such projects for SMBs. Through the conclusions and recommendations from the research, the SMBs can assess data analytics implementation possibilities and minimise the likelihood of failure during the implementation process.
1.4.2 Thesis Structure
The thesis structure provides a gradual transition of theoretical foundations and findings to their practical implementation for the particular cases. The paper consists of 10 chapters that can be grouped into four parts.
The first part describes the theoretical foundations of Data analytics and its usage by SMBs and information about the research aim, objectives, structure and methods used. The description of the case companies also provided there.
The following part relates to the theoretical review of implementation methodologies and approaches with the assessment of the companies maturity levels. Within this part, data analytics projects for case companies are formulated and the key structured elements of data analyses in the company were investigated.
The next part relates to the analyses of the data analytics ecosystem and solutions on the market that can be used by SMBs with relation to the cost and benefits analyses. Here the current data analytics architecture of case companies was provided.
The final part is the creation of the decision-making model and its application to case companies.
1.4.3 Research Limitations
The research shall be completed in one year. Due to the time limit, the suggested decision- making model could be tested on a limited number of cases without implementation of the selected solutions.
1.4.4 Methods description
The research assumes applying a combination of several qualitative research methods. The usage of qualitative methods is driven by the need for a deep understanding of the problem and analysis of specific cases and suits for creating new approaches and solutions. The first method is secondary research by a literature review to collect and analyse the current developments in the data analytics sphere. The second is the case study that is the method for an in-depth examination (Saldana, Leavy and Beretvas, 2011). Gummesson (Gummesson, 2017) gives the following guide of the research journey that was applied in the research:
1. Research plan. At this stage, the problem, research questions and purpose should be defined, the access to case data theory selected, analysis, interpretation and conclusions framework designed, determined the way of reporting and the target group for it.
17 2. Access to case data. To provide the correct data, a combination of methods can be used. They should be selected according to the specifics of the data needed to be collected. In the research, interviews with the employees of the case companies are conducted. It should be noted that data collection time plays a crucial role because data collected at the beginning of the research could become out of date at the time of analysis. For this reason, two case companies were investigated for 4 months in parallel.
3. Analysis and interpretation. The idea is to link data, theory and conclusions by step by step process of extracting information and knowledge from data to generate theory and conclusion. Within the framework of this study the steady analyses of the business, data analytics structure and elements are conducted. The final solutions and recommendation of the research base on the theoretical study of the current data analytics solutions and cost/benefits evaluation techniques.
4. Research quality and productivity. Those concepts are used for the evaluation of the research process and results.
5. Report, communicate and defend results to the target group. It includes the final presentation of the results to the decision-makers, discussion of the alternatives and final decision.
Case 1: company description
The first case company is the supplies high-quality petroleum products, operating in a small town in a CIS European country. The company sells liquefied gases, fuel oil, road bitumen, diesel fuel, unleaded gasoline domestically, and export. Unfortunately, the company was not allowed to be named in the study because of data sensitivity and the specifics of the market.
Due to this reason, the company will be named as Case 1 throughout the work.
The company is a typical representative of a medium-sized business. It has a Limited Trade Development form of ownership. It does not have representative offices, trade, and other organisations established abroad and does not receive support from the state. The primary activity of the Case 1 is wholesale trade in petroleum products. However, the company provides other services. According to the results of financial and economic activities for 2020, the revenue from the sale of products, work performed, services provided amounted to 9,5 mln euro. In 2020 the structure of revenue from sales was as follow:
wholesale of petroleum products for road construction companies - 58%;
retail sales of the petroleum products - 40%. The sale is carried out through 3 petrol stations that the company owns;
proceeds from the provision of customs clearance services - 2%.
The company is also engaged in the export of petroleum products, which account for 30% of its total revenues.
At the end of 2020, the total number of employees was 77 people. The organisation structure (Figure 2) includes seven departments and four specialists who report directly to the CEO.
Each of the company’s activity has a separate department: Sales department responsible for
18 wholesale, Custom clearance work with clearance services and Petrol station management organise the work of retail sales.
Figure 2. Organisation structure of the company, Case 1 (Author, 2020)
Due to the fact, the company is LTD, the CEO is responsible to the supervisory board who accept the company's strategic goals and activities, including the budget plan. To ensure a stable work of the company and reduce mistakes in working with high priced products, the own IS/ICT structure was created, that includes software and hardware for automation of the work of the production department of customs clearance, for obtaining reference and legal information, automation of accounting and management (including payroll and personnel management), electronic document management systems. The company pays a lot of attention to the level company equipped technically as one of the key factors of the company's effective work. The systems in use are regularly updated, new ones are purchased, outdated computer equipment is replaced with a new one as needed. The total costs spent on ICT in 2020 amounted to about 26 thousand euros.
Case 2: company description
The second company included in the case study is the producer of the packaging, operating in the capital in a CIS European country. To separate two cases, the name “Case 2” will be used in the research. The company creates different kinds of boxes using corrugated cardboard and micro-corrugated. The production process is full-cycle which means it constructs the packaging starting from design, selecting material and ending with offset printing. Average monthly production volumes are 100-120 thousand m2 and turnover varies between 90 - 150 thousand euros.
19 Figure 3. Organisation structure of the company, Case 2 (Author, 2020)
According to the organisation structure (Figure 3), the company has five structural units. The number of employees includes 40 full-time employed staff and 5-10 employees working permanently under a contract. The commercial department includes sales specialists and finances. Engineering provides adjustment and installation of equipment. The planned technical department is responsible for producing and shipment of products and contains technologists, engineers and designers. The pre-production department is responsible for the colour solution and preparation of the systems for production. The company has two production lines, and each of them is also equipped with employees to control the production process.
20
2 Data analytics implementation literature review
Any implementation and development activity in the company is a project that covers all the stages from its initialisation to its closure. The decision-making process is a part of the initialisation phase, where the evaluation of possible variants of the realisation is available and the decision about the final decision is approved. This stage answers the questions about feasibility (“Is it possible to do the project?”) and justification (“Should we do it?”) (Watt, 2014).
To ensure a steady and smooth development, different methodologies, models and frameworks were discovered. One of the groups refers to the data mining and knowledge discovery process.
Mariscal, Marbán and Fernández (Mariscal, Marbán and Fernández, 2010) provided a comparison of 15 methodologies and models from this field. According to the results of the research, all of the methodologies have an analytical stage before the implementation process.
However, each of the methodologies includes different activities in this stage. The summary of activities in the analytical stage is: life cycle selection, domain knowledge elicitation, human resource identification, problem specification, data prospecting and data cleaning.
In addition, it was found that the majority of these approaches are based on two main methodologies - KDD and CRISP-DM. CRISP-DM methodology is one of the most used today (Piatetsky, no date) and well developed (Chapman et al., 1999).
CRISP - DM methodology includes a life cycle of 6 steps: business understanding, data preparation, modelling, evaluation, deployment. For the purpose of the current research, the first phase of the CRISP-DM lifecycle about business understanding is of interest. This stage is the crucial difference of this model from the older one, KDD. There are four substages inside this stage:
Determining business objectives. It includes research of the background, understanding business objectives, choosing business success criteria.
Assess situation. At this stage, the inventory of the existing sources provided, defining the requirements, list of assumptions and constraints, risks, costs and benefits.
Determining data mining goals and success criteria.
Produce project plan and initial assessment of tools and techniques.
CRISP-DM model still does not lose the leading position; however, this model does not fully meet the requirements of the current level of development of analytics and technology.
Following limitations can be defined for this model:
lack of focus on the human role (Rotondo and Quilligan, 2020);
waterfall life cycle with limited iterations or interactions between phases (Rotondo and Quilligan, 2020);
it does not cover tasks related to project management, organisation and quality management (Mariscal, Marbán and Fernández, 2010);
lack of consideration of various data sources (Rotondo and Quilligan, 2020);
21
lack of integration of software development processes (Rotondo and Quilligan, 2020);
not maintained and adapted to the challenges of modern data science technologies (Piatetsky, no date).
Another group of the methodologies and frameworks relates to the Business Intelligence implementation (Loshin, 2012; Sherman, 2014; Data-driven business transformation, 2015), including also methodologies for implementation in SMBs (Raymond, 2003; Guarda et al., 2013). The authors provide detailed descriptions of the phenomenon, its architecture and design with the relation to the planning process, creation of strategy and requirements to BI projects. Lila Rao-Graham, Maurice L. McNaughton, and Gunjan Mansingh(Rao-Graham, McNaughton and Mansingh, 2019) provided their BI roadmap development plan with the focus on SMBs and agile methodology. The approach was developed as opposed to the CRISP- DM methodology. Despite the fact that the model is designed for BI implementation, it also includes a prediction layer and can be used for advanced analytics implementation too.
The model draws attention to the fact that according to the CRISP-DM model, companies are already ready to implement analytical tools and techniques when in reality, not preparing for such investments leads to project failure. AIMS-BI provides a methodology to assess the company’s readiness, identify the capability gaps and develop a strategic path in an agile way (Rao-Graham, McNaughton and Mansingh, 2019).
This methodology includes four steps before creating a strategic roadmap (Figure 4). The first step is an information maturity assessment of the companies. The survey and capability assessment are used here for an analysis of the maturity baseline and existing gaps. The second stage is the discovery of BI opportunities by conducting interviews and case studies. At this stage, the understanding of the business provided, highlights the strategic priorities and opportunities. The third stage is the BI portfolio evaluation, where the initiatives are evaluated and priorities selected. The fourth stage is the Proof of concept, where the prototype is analysing, data and process management established and evaluated. The result of all these stages and the methodology’s output is the Strategic roadmap that derives an appropriate configuration according to the objectives and circumstances (Rao-Graham, McNaughton and Mansingh, 2019).
Figure 4. AIMS-BI methodology (Rao-Graham, McNaughton and Mansingh, 2019)
22 Loshin (Loshin, 2012) focuses on the following aspects of the BI project initiation:
Championship where the corporate sponsorship is ensuring.
Level-Setting with defining goals and expectations.
Partnering among participants to ensure acting strategically.
Vision asserting that influence on design, development and deployment of the project.
Plan development with the set of tactics that ensure achieving the long-term goals.
Sherman (Sherman, 2014) provided another guidance for BI project preparation. He identified 3 phases of roadmap development: discovery, analysis and recommendations. In the first phase, the business and IT requirements, organisation, partnership, process, skills, data, information, technology and product architecture should be assessed. The result is an understanding of the current and desired reporting and analysis, priorities and analytical skills.
In the analytical phase, the gap between the desired and current is examined based on the information from the previous stage (architecture, organisation and requirements) and validation of the ability to achieve future end state.
The recommendation stage is a finalisation of the investigations during the analytics stage and providing a road map. These recommendations will involve architecture (information, data, technology, and product) and organisation (program/project plans, resources/skills/training, and rough cost estimates.)
The final solution lists should be followed by the following attributes (Sherman, 2014):
Key business initiatives supported and business processes involved.
High-level business deliverables.
Data source feeds (at the application level, not tables or files).
Technology introduced.
Business groups involved.
Finally, the roadmap is created according to the plan that details the resources, tasks, and schedule. This information is used for analysis, design, validation of the design, development, implementation, and deployment process.
A detailed explanation of the data analytics implementation and governance in the company was provided by Sheikh (Sheikh, 2013) and Slansky (Slánský, 2018b). Sheikh (Sheikh, 2013) provided guidance for plan, design and build analytics solution and introduced his Information continuum as a basis of his approach. The idea of the continuum is the steady development of data analytics from a simple search of data generated by different sources to future projection and decision-making optimisation. Monitoring and turnover referred to the governance of analytical models and strategies and used all previous levels to ensure that the analytical model and decisions perform correctly. However, the disadvantage of the approach is the focus on the data mining process and traditional technologies. The project lifecycle focused on the data mining processes model and based on traditional ETL and data warehouse solutions that may not always be suitable for SMBs (Rao-Graham, McNaughton and Mansingh, 2019).
23 Slansky (Slánský, 2018b) provided a broad description of data and analytics architecture with the components and the guidance of their governance. There are also practical tools for components evaluation and implementation. He defined seven areas of data analytics governance that covers the development and operation lifecycles: strategy, understanding, requirements management, development management, architecture management, operations management, budgeting and specific areas (data quality management and data security management). All the areas were described from the point of vision, objects, processes, tools, organisation structure and metrics used in particular areas. However, the work focuses more on the architecture and governance of the data analytics than on the step by step plan of its development.
The question about preparation for data analytics implementation was also developed by the companies that make such implementations or provide consulting services. Each company develops its own approach of evaluation and analysis that they use in practice. However, because these developments represent a company’s unique development, the information is not freely available. It is possible to find only description of some parts of the process. In addition, even those parts do not have a detailed explanation about the process and tools companies used on each process stage. Intel (Intel Corporation, 2015) uses 5 step planning activity and provided a decision tree for finding a suitable data analytics solution based on technological and organisational factors. IBM (‘IBM Data and Analytics Strategy Field Guide’, 2018) shared their data and analytics strategy creation approach to creating a roadmap for small businesses.
Gartner (‘The IT Roadmap for Data and Analytics’, 2020) provide a brief description of the roadmap with key objectives and outcomes on each stage. However, the list is without complete details of all milestones and resources for key steps. Deloitte (Data Analytics, no date) mentioned the idea of data analytics strategy that include fundamental blocks (people, processes, technologies and data) and a three-stage approach of data analytics adoption (assessment, roadmap, and deliver). Thus, the complete process of analysis for data analytics implementation, including the analytical stage, provided by commercial companies is not available.
It should be noted that there is a high interest in applying agile methodologies in the analytical and development process. In addition to the already mentioned works of Lila Rao-Graham, Maurice L. McNaughton, and Gunjan Mansingh (Rao-Graham, McNaughton and Mansingh, 2019), it was also used by Deanne Larson and Victor Chang (Larson and Chang, 2016), Gartner (Parenteau et al., 2015) and Eckerson Group (Wells and May, 2017).
24
3 Company's Data analytics maturity assessment
To access the company’s current situation regarding data analytics, the evaluation of organisations’ maturity level should be provided. Such evaluation should be done before the start of the project and after its execution to demonstrate progress (Jackson and Carruthers, 2019). The analysis of the company’s place in the maturity hierarchy can help to understand tools, skills, infrastructure, and other elements needed and how to deal with models and decision strategies (Sheikh, 2013).
To determine the level of maturity in the company, different maturity models were executed.
They help organisations to find out their strengths and weaknesses and determine the further direction of development. According to the analysis of different data analytics maturity models, Krol and Zdonek (Król and Zdonek, 2020) determined that all the models share a common identification of 3 factors that influence the analytics development: human resources, infrastructure (equipment and software) and resource management. As a result, the assessment of the maturity level primarily based on these dimensions and include:
an evaluation of technical infrastructure (the equipment and software, data collection methods);
an assessment of organisational issues (analytics culture, the degree of support and democratisation of analytics, and the level of acceptance towards an analytics culture in the entire enterprise) ;
an assessment of human resources (the staff’s analytics competencies).
All the models have very similar dividing of the maturity stages. The models of Gartner (White and Oestreich, 2017), Peter Jackson , and Caroline Carruthers (Jackson and Carruthers, 2019), Rao Graham (Rao-Graham, McNaughton and Mansingh, 2019) has next five levels of maturity:
Level 1- Aware (or Basic). There is no standard practice in data analytics management. The existing processes are required by legislation or industry regulations.
Mostly, they are ad hoc efforts and with low levels of trust. Some examples of higher levels can appear only in the case of the initiatives of individuals.
Level 2 - Reactive (or Opportunistic). The practices of data and information management used a little and mostly in specific departments. The data becomes a little bit insightful, however still not fully organised. When some processes emerge, they are only monitored regularly, and data quality checked ad hoc. There are no standards for such processes and response is only taking place when the issue emerges. Even when IT tries to formalise the information availability requirements, the culture is not ready and does not allow progress.
Level 3 - Proactive (or Systematic). The stage where the data management programmes and unstructured data appeared. However, content can be still treated differently. The policies, procedures and standards exist and used across the
25 enterprise. The data and information management practices are typically sponsored and managed by IT using the agile approach.
Level 4 - Managed (or Differentiating). Information governance is in place and performs as an enterprise asset. Data analytics become a source of performance and innovations. A lot of matured best practices are developed, and audit is used to ensure compliance across all projects.
Level 5 - Optimized (or Transformation). Information is the same asset for the company as financial or material assets. Almost all data assets are inventoried, including external sources and information distributed throughout the business to create value. The best practices used across the whole company and focus on continuous improvement. The methods of data management practices and assets are regularly measured, analysed and improved. Heudecker (Gartner Survey Shows Organizations Are Slow to Advance in Data and Analytics, 2018) claims that organisations in transformation use all the positive impacts from agility, better integration with partners and suppliers, and easily implement advanced predictive and prescriptive.
Peter Jackson and Caroline Carruthers (Jackson and Carruthers, 2019) also have in their model level 0. Usually, companies do not have this level; however, it can be possible that a company does not care about the data it produces. Each employee works with his own Excel, taking data from different sources and modifying it without understanding that it can be helpful for others.
To provide a maturity evaluation for the research companies, the questionnaire provided by Peter Jackson and Caroline Carruthers (Jackson and Carruthers, 2019) was used. The list of questions was optional, and it was slightly adapted to the specifics of SMBs for this research.
The analysis was done in 11 dimensions: Strategy, Corporate Governance, Leadership, Framework, Policies, Risk, Architecture, Organisation, Skills, Metrics, and Behaviour. The survey took place within a semi-structured interview framework, where some questions could be swapped or not asked. The list of questions could be found in Attachment 1.
Based on the analysis, the company from Case 1 can be attributed to the second proactive maturity level. The company does not work with unstructured data, and there is no established data strategy. The process of processing and organising information is adjusted to specific requests. However, the company knows how they use and treat data, has a clear structure of data governance that all the responsible employees know. Using the data is critical for the company's performance, and the leadership understands the importance of managing data appropriately. The management of the data is a part of the information system management.
For this reason, the sponsorship and prioritisation relate to the whole information system in the company. Having a clear information system structure helps the company to ensure the understanding of data sources and their interconnection, their regular maintenance according to the information lifecycle. The company has clearly defined roles and enough skills to manage at this level. There are documented policies about data control; however, they are not clear for understanding. The company understands the risks related to data security and transformation in different systems. For this reason, additional processes for checking data correctness were established. However, the maturity of regulations, structure and risks evaluation is different depending on the department. The reason is that each activity of the
26 company works separately, has their own processes of information management and measurements of the results.
The company from Case 2 belongs to the first level of maturity. The company does not have an established data strategy and corporate governance that can regulate the data management and analytics processes. However, the management understands the value of the data and its analysis for the company's performance and evaluates the priority of this as middle high. The company has a clear structure of data sources that update mostly manually every day and have all actual information. The data is also processed according to the established rules; however, these rules are not documented. The general architecture is clear for every employee and responsible person and their tasks are defined. However, there is no certainty that the company has enough skills to work with data the right way. The metrics the company uses are basics, the majority of the reports required by legislation. They also use ad-hoc reports that prepare ones in decade or year. If to speak about the risks, the company does not focus on this question and does not manage them, because they are not working with sensitive data. However, they assume that there can be some hidden risks that they are not taking into consideration.
27
4 Business understanding and project formulation
To understand business goals and formulate the data analytics project’s idea, semi structured interviews with the companies’ CEOs and employees were conducted. According to the analysis of the studies, the answers to the following questions should be found during the interview:
1. What are the business goals and objectives (Intel Corporation, 2015; Rao-Graham, McNaughton and Mansingh, 2019).
2. What are the business problems to be solved using data analytics and business requirements (Chapman et al., 1999; Sherman, 2014; Intel Corporation, 2015) .
3. What is currently used from data analytics, the advantages and disadvantages (Chapman et al., 1999; Sherman, 2014; Rao-Graham, McNaughton and Mansingh, 2019) .
4. Formulate the projects and evaluate which of them are more likely to be done according to their impact and expected ROI (Intel Corporation, 2015; Rao-Graham, McNaughton and Mansingh, 2019) .
5. Which business processes the projects covers or make impact on (Sherman, 2014; Rao- Graham, McNaughton and Mansingh, 2019) .
6. Who are the target groups and users (Chapman et al., 1999; Sherman, 2014; Intel Corporation, 2015; Rao-Graham, McNaughton and Mansingh, 2019) .
7. What are the companies needs and benefits the company expects (Chapman et al., 1999; Sherman, 2014; Rao-Graham, McNaughton and Mansingh, 2019) .
As for any other commercial company, the key goal for Case 1 is increasing profit. According to the plans the CEO provided, the company is going to develop each of the activities of the company. The first is an optimisation of the wholesale of petroleum products by improving the form of settlements with suppliers and creditors and eliminating the impact of changes in the currency exchange rate. With increasing uncertainty in the market, an accurate analysis of the effects of such changes becomes critical.
In the case of the retail trade, it is planned to reconstruct petrol stations and optimise the range of products sold. Selling related products is the most profitable part of this activity and should be optimised to maximise profits. The last focus is the development of the provision of customs clearance services. The plan is to expand the types of work within the service and attract new customers by this.
The realisation of these plans requires instruments for making decisions and tracking the effectiveness of implemented activities. Currently, the company collects a lot of financial data that they use for both operational and strategic decisions. The information about the financial activity is created by the accounting department and analysed by economists regularly. The sales department uses information about currency rate volatility and price changes in the market to manage bitumen supplies. Petrol station management analyses the product assortment and leftovers to determine the volume of new purchases of products. The information the company uses is enough for making decisions; however, they see the potential to speed up and improve the analysis’s accuracy.
28 After analysing the current activities of the company and discussion with the companies’ CEOs, the following projects for data analytics development were identified:
1. Informing about the company’s current performance for each activity in real time and showing future income and expenses and overdue payments.
This form of reporting already exists and is prepared by economists; however, these reports need a long time and a lot of human resources for its preparation. Optimisation of such reports will help economists and CEO react faster to negative changes in the company and also anticipate them in advance.
2. Analysis of the delivery of related products to petrol stations for sale.
Such analysis will help managers predict and select the range of goods and their volumes more accurately based on the day of the week, seasonality, and stock balances. As a result, it will increase sales and optimise the costs of purchasing and storing products. Currently, daily purchases are carried out based on product leftovers and employees’ experience on previous purchases.
The declaration service is in development now. That is why some extra analytical solutions except financial analysis that already exist in the company do not seem necessary. All mentioned projects have a high value on the company performance. However, the first project is critical for the whole company and creates value for each activity the company performs.
In the case of the second company, the CEO works with two key objectives:
increasing the number of clients by attracting new clients from the SMB sector inside the country;
optimising the cash flow to avoid cash gaps.
These objectives appeared due to the economic crisis that decreased the number of companies that require the production of packaging and the average order volume of key clients. Due to the reason packaging is a seasonal product and seasons do not coincide for different industries;
changes in the customer portfolio lead to a significant difference in cash flow in different months, including the appearance of a negative balance at the end of the month. It is also influenced by the time and volume to order materials. The order should be done with minimal extra costs and low prices. However, it is difficult to accomplish with decreased volume production. In this situation, the warehouse management with remnants of materials and ready products can play a crucial role in reducing unnecessary costs.
Currently, the CEO, heads of the Commercial, Accounting, Planning and Despatching, and Production Preparation departments use data for decision-making in their work. The analysis covers the needs of sales, production and finance objectives. The purposes of these analyses are:
assessment of the future production and materials required based on the customer order;
analysis of the company's financial performance.
29 However, in a rapidly changing environment, current analytics outputs are not enough to make data-driven decisions on the new appeared issues. After the discussion of the current state, the following analytics projects was identified:
1. Provide information about clients who send requests through the website.
It will help create the profile of a potential client, understand popular queries in the market, find new segments for targeting, and adapt product portfolio. The output will be used by the Commercial director and CEO for improvement in the process of attracting potential clients, sales and production.
2. Create charts with information about current orders and payments of customers, possible future orders (for regular clients), and a seasonality chart of the markets with which the company works.
With the output from this solution, companies can concentrate on the markets they need to eliminate a cash gap, predict profit, and find suitable clients. This information will be helpful for the Commercial Director and CEO. It covers the sales and procurement process.
3. Providing the accounting of balances of materials and products in the warehouse in real-time with the required volume of materials for existing orders.
Working with such information will increase the quality of work planning, accurate purchase and reduce costs. It will be valuable for the Production preparation department.
For now, this information is created each month manually for the procurement process and updated in case of a new order.
The most valuable project here is the second one because it covers all three issues the company has in hand and can provide a fast outcome. However, the realisation of all three projects can optimise the entire production process from purchasing materials to the release of finished high-quality products and their shipment to the customer. This will allow the company to accurately determine product readiness and delivery, which is the crucial competitive advantage.
30
5 Analysis of Data analytics structure elements
Data analytics require core structure elements in the company to be provided. They create a business ecosystem triangle where the data is a driver behind transformation (Jackson and Carruthers, 2019). These elements are data, people, tools and processes (Figure 5) (Lawson, Hatch and Desroches, 2019). Assessment and inventory of the key elements give a clear understanding of the current state of data analysis in the company and it’s potential.
Figure 5. Business and data ecosystem transformation (Jackson and Carruthers, 2019)
5.1 State-of-the-art tools and techniques
To transform data into understandable insights, data analytics tools and techniques are used.
More than half of the SMBs claim that the biggest problem they face is to choose between different solutions and the lack of experience to evaluate them (Data Analytics Trends, 2020).
On a basic level, companies use BI platforms embedded in ERP systems or other simple reporting tools. A more advanced level implies the creation of integrated analytics platforms with modern analytical technologies (Moore, 2018). It should be also taken into account that the industry where the organisation act, its regulations, technological changes, standards can shape what technologies and how is used within the company to create business value (Melville, Kraemer, and Gurbaxani, 2004).
Loshin (Loshin, 2012) notes that a company needs to allow innovations as long as it creates potential value for the company that can be measured. To do this, they need to identify successful techniques and ensure that they have processes to integrate them into enterprise architecture. However, Microsoft Excel is still the most commonly used reporting and analysis tool despite its limitation with sharing and changing the outputs (Slánský, 2018a). The Gartner research (Gartner Survey Shows Organizations Are Slow to Advance in Data and Analytics,
