Abstract

Business intelligence (BI) is an architecture intended to gather, explore, analyze, mine and provide access to heterogeneous sources of business data. The article aims to define Business Intelligence, describe the difference between on-premises and cloud BI solutions. Most importantly, the thesis focuses on the identification of selection criteria for those different BI solutions, as a way to help business managers and IT leaders to select tools which fit best with their specific requirements. Although Business Intelligence is widely used in industry, frameworks for choosing a business intelligence tool are rather limited. To do so, the thesis relies on a review of potential selection criteria for BI solutions, and on the empirical evaluation of those criteria based on data gathered from different profiles of BI specialists.

Keywords: Business Intelligence, BI, on-premises, cloud, criteria selection, quantitative analysis

1         Introduction

Nowadays, it is important for every company to not only keep its level of competitiveness, but also to increase it. Information technologies are developing at a high pace. The need for flexible access to different types of data and a complex analysis of data is becoming a necessity for every business.

The information systems market offers a variety of choices of solutions that helps organizations to store and analyze information, facilitates decision making in the long-run and stables results. An important part of this market is occupied by Business Intelligence solutions.

Business intelligence (BI) are tools that help users to select the information about the organizations and its surroundings. BI technologies help to treat large amounts of data, focusing only on the key factors and monitoring the results to help in decision-making.

Business Intelligence helps in converting data into knowledge and knowledge in its turn can be transformed into action, which can mean profit for the business.

The aim of this article is to answer the research question which is focusing on the criteria that matter when an organization wishes to implement a new on-premises or cloud BI tool.

In order to meet the goal set above, the following needs to done:

1 – Study the concept of Business Intelligence

2 – Define the different types of Business Intelligence used in the industry

3 – Make an overview of tools and where they are used

4 – Conduct a study on the criteria that matter for choosing a Business Intelligence tool

The practical implication of the study is the ability for organizations to use a framework that can help them to choose the correct BI tool for their organization.

2         Literature

Across every industry, organizations are on a road to putting data at the centre of business transformation, whether the goal is to better understand customers, build new or better products and services, or to manage costs and risks. (Cloudera, Inc, 2014). Nowadays, information and knowledge represent the fundamental wealth of an organization. Enterprises try to utilize this wealth to gain competitive advantage when making important decisions. (Ghazanfari, Jafari, & Rouhani, 2011) The key is to extract data that the business requires out of a well-documented data landscape, determine the data requirements that will affect the success of the chosen strategy, then profile and correct the data so it is useful for business needs that the later finds critical for its success. (Ladley, 2016)

Data warehouse

According to The Data Warehouse Institute, a data warehouse is the foundation for a successful BI program. It a repository for every company’s data. The data warehouse (Figure 1) gets data from different kind of sources, including the external sources, extraction, transformation and loading (ETL) or from online transaction processing databases (OLTP). When the data is gathered, the BI tools can be used in order to analyze the information.

Figure1. Relationship between BI and Data warehousing. Source: Poolet (2007)

When it is needed to design a data warehouse, there are 2 approached that are commonly cited that can be used, the ones of Bill Inmon and Ralph Kimball.

• Bill Inmon’s approach (the top-down design): a normalized data model is created first and the specific data marts are added. This approach minimize data redundancy, as the information from the data model is linked to the data marts in order to created different tables and connections. (Inmon, 2002)

• Ralph Kimball’s approach (the bottom-up design): every data mart is created first which are later combined in order to create a data warehouse. In Kimball’s approach the data is organized following a dimensional model, which consist of a fact table that is then linked to different dimensions. The dimensions data is de-normalized.  (Kimball, Ross, Mundy, Thornthwaite, & Becker, 2011)

After a data warehouse is designed the BI software should be chosen. Analysts from BARC, a leading enterprise software industry firm, list the 4 essential steps in the BI selection process. They point out that those 4 steps are essential for a successful fit of a BI tool within an organization:

Figure 2: 4 essential steps in BI selection process. Source: BI Survey (2016)

1. Strategy/Goal definition: Before the implementation of any kind of BI systems within an organization, it is firstly needed to have an overall vision of the system and kind of value can it bring to the organization. It is important to understand that a company that wishes to implement a BI tool should not only focus on data collection, but the goal should be to bring value from interpreting the collected data. It should be aligned with corporate objectives of the organization. (Olzak & Ziemba, 2007)
2. Requirements Analysis: at this stage it is important to analyze and describe the logical structure of data to be found in the system. The organization should determine what does it expect from the system and those expectations should be quantifiable and detailed. (Dresner, et al., 2002)

3. Software Evaluation: the offer of BI tools in the market is rather important and the selection of a tool may be a rather difficult task. In order to select the correct tool, a wide range of criteria should be considered and a framework that can help in choosing a tool might be helpful for an organization.
4. Implementation and deployment: this stage of the process focuses on making sure that the BI application is logical and running smoothly. It also involves testing and the making of changes, if necessary.

What is Business Intelligence?

According to Cebotarean (Cebotarean , 2011), business intelligence (BI) refers to computer-based techniques used in spotting, digging-out, and analyzing business data. BI technologies provide historical, current, and predictive views of business operations. BI systems are now used extensively in many areas of business that involve making decisions to create value. BI systems contribute to improvement and transparency of information flows and knowledge management (Kalakota, & Robinson, 1999; Liautaud, & Hammond, 2002; Moss, & Alert, 2003) However, to help BI achieve its full potential, we need to fully understand the processes through which organizations can get value from BI. (Trieu, 2016)

Gartner defines cloud computing as a style of computing where massively scalable IT-enabled capabilities are delivered as a service to external customers using the Internet. “The cloud has been a hot topic for a number of years, with companies moving applications to the cloud for speed to execution, lower costs, and higher level of service and/or preservation of capital”, writes Marc Malizia, CTO of RKON Technologies, a managed cloud solutions provider.Cloud BI, which hosts the software on the Internet, reduce storage costs and make access to organizational data and insights faster and more convenient (Heinze, 2014).

“Cloud computing and business intelligence are an ideal match”, mentions Klipfolio (2017). Business intelligence is about helping the user to understand the data in a shorter amount of time, and cloud computing a quick and easy way of accessing the tool.

Nowadays BI is (almost) analyst driven, as we have arrived to the so-called Big Data movement. The big data revolution and explosion of the Internet left organizations with more data than before. Visualization tools began to evolve to include the end-user even more. More platforms empowered users to complete self-service access, meaning that they could explore and utilize their data on their own, without training (Heinze, 2014). According to a Gartner analyst Doug Laney (2001), Big Data is high-volume, high-velocity and/or high-variety information assets that demand cost-effective, innovative forms of information processing that enable enhanced insight, decision making, and process automation. A recent article by Brown (2015) describes the difference between the BI and big data:” Business Intelligence helps find answers to questions you know. Big Data helps you find the questions you don’t know you want to ask”. The Big Data concept will not be discussed in this article.

The figure 1 below shows the year over year evolution of different BI technologies (mobile and social BI are not part of the research and will not be discussed in the article). The year 2013 serves as a base year and is not represented on the graph. It can be seen that year over year, different technologies emerge and take their place on the market.

Figure 1. Global business intelligence market by technologies, 2013-2018. Source: (Gartner, Redwood Capital).

Difference between on-premises and cloud BI

Many BI projects require traditional BI implementation. These projects typically require real-time data and connectivity to many different data sources. But there are many other BI projects that require neither real-time data nor the data which comes from different systems that should be connected. Cloud BI can handle these projects quickly and cost effectively, by empowering business users to manage the whole process without IT or external support

A recent survey (Gooddata, 2016) indicates that on-premises (traditional) BI systems are neither flexible nor agile enough to keep pace with online sources that churn out data by the second. The value of real-time data has become clear, forcing BI systems to evolve from complex infrastructure investments toward more agile, cloud-based platforms. Cloud BI is more flexible, easier to scale and, most important, much faster to deploy than ever before. As a result, larger and more obscure data sets can be collected and turned into a jaw- dropping array of raw digital inputs from nearly every internal and external aspect of business.

Overview of tools and where they are used

There is a large number of different BI tools on the market. BARC (2017) made a user review matrix to present the customer experience versus the business value ratio for large international vendors. As we can see form the bellow graph (figure 3), there is an important number of tools in the top right corner, which represent the maximum business value and user experience. Some of the below tools are present in the traditional version of a BI toll as well as in the cloud version. Companies, that develop BI tools tend to develop traditional and cloud solutions more and more, in order to fit the demand of large companies as well as SMB’  (Small and medium businesses).

Figure 3: BI vendors. Source: BARC(2017)

Gartner, an information technology research organization, developed the concept of magical quadrant (Figure 4), which differs from the BARC BI user review matrix.  We can see 4 types of players in the magic quadrant: the nice players (focusing on small segment or do not perform well), the visionaries (have a good understanding of the market, but do not execute well), the challengers (execute well, but do not have a complete understanding of the market of tomorrow) and the leaders (execute well today & potentially tomorrow).

Figure 4. BI Magic Quadrant. Source: Gartner (2017)

According to Willen (2002), BI assists in strategic decision making. The strategic use of BI is characterized by the following:

• Corporate performance management
• Optimizing customer relations, monitoring business activity and traditional decision support
• Packaged standalone BI applications for specific operations or strategies
• Management reporting of business intelligence

Business intelligence tools are used in a variety of ways. BI survey (2016) made an analysis of the widest tasks where BI used in the below graph.

Figure 4: Tasks where BI is used. (Business Application Research Center, 2017)

We can see that companies tend to mostly use BI for the following 4 tasks:

Standardreporting: collecting data from various data sources and presenting it to end-users in a way that is understandable and ready to be analyzed (Logi Analytics, 2017)

Ad–hocquery: according to Merriam-Webster Dictionary, ad hoc means “for the particular case at hand without consideration of wider application.” The ad-hoc queries are used when a user needs a deeper understanding of a particular standard report.

Basicdataanalytics:  according to Gartner, it is “a statistical and mathematical data analysis that clusters, segments, scores and predicts what scenarios are most likely to happen”.

Dashboards: a combination of data that is analyzed by the BI tool and that shows as an output a consolidated view of metrics and performance of the organization on a single screen. (Hart, 2017)

Criteria for choosing a BI tool

When making a choice on the tool an enterprise would like to use for its business, criterions on which the decision will be made should be chosen. Many different sources list a large number of criteria that are important when an enterprise wants to implement a BI tool into the company.

BARC (2010) lists the criteria based on which the software should be chosen and how to companies actually choose it: