In this excerpt from the book Using Open Source Platforms for Business Intelligence: Avoid Pitfalls and Maximize ROI, author Lyndsay Wise, president and founder of Toronto-based consultancy WiseAnalytics, digs into the topic of BI adoption. Wise examines why organizations might consider implementing business intelligence software in general and open source BI tools in particular. She looks at how businesses can benefit from a BI strategy but discusses the importance of assessing business goals and existing IT infrastructure before starting a deployment.
The first of three segments from Chapter 8 of Wise's book, this section introduces key factors that companies must consider in weighing a choice between open source business intelligence software and conventional BI technology. In addition, readers will learn about some of the advantages that result from adopting BI tools, including data consolidation, increased organizational visibility and improved risk mitigation.
The strategy behind BI adoption
Understanding the world of open source business intelligence (OSBI), some of the differences within the marketplace, and why vendors position themselves as they do are a first step towards identifying whether this type of BI deployment matches the requirements of your organization. If you are looking for an OSBI offering or trying to decide the type of deployment you should select, the next set of questions you and your colleagues should be asking are "what do I do with this information?" And "how do I decide what will work best for the organization?" After all, software selection is not as easy as closing your eyes and pointing to a vendor's logo on your computer screen. It requires being able to sell the value of BI to management and defining what isn't working in the current system, how BI can make it better, and what options might work best within your IT environment.
This chapter has the lofty goal of addressing these questions. After all, in order to identify whether OSBI is right for you, it is important to first consider how BI will support the organization's strategic goals. Once you make this decision, it becomes possible to identify whether OS fits within the company's business strategy. Obviously, some organizations bypass in-depth evaluations and select software based on previous use or recommendations from friends. Although this can also work, the greater the understanding of what exists internally, the more likely that the overall project will be successful.
One of the problems with the BI marketplace today is that unless an organization is well versed in all of the solutions available, most solutions look alike and can provide similar things at face value. Yes, vendors have their key differentiations, but the reality is that most solutions can be customized to provide like features and functionality. Being able to optimize the solution, however, is one of the key differences when going the distance and taking the time to match vendor requirements with business goals and the existing IT infrastructure.
This excerpt is from the book Using Open Source Platforms for Business Intelligence: Avoid Pitfalls and Maximize ROI, by Lyndsay Wise. Published by Morgan Kaufmann Publishers, Burlington, Mass. ISBN 978-0124158115. Copyright 2012, Elsevier BV. For more info, please visit the Elsevier store website.
Let's take this one step further by looking at the role of developers and the process they need to go through to implement an end user facing BI solution. Generally, IT developers take a lot of time putting the pieces of the data puzzle together. This means they look at each table within each data source and identify if and how information is related and which fields are required. For instance, what product information exists and where it is stored, how it relates in one table to various other tables, what information is required, and what calculations are required to make sure that business users can get the insights they require to make the best decisions possible.
Doing all of this effectively is no easy task. Some tables may house Product Number, while others use a separate Product ID descriptor, with others still providing a Name and Description. To create a singular view of data, all of this information needs to be consolidated and, in many cases, product numbers do not follow a specific structure. So, naming conventions might not be standardized and like names or descriptions might be used that mean different things to different employees. And even within data warehouse infrastructures that pull in all of the data without all of these data joins and additional preparation, these tasks are still required to develop strategic queries.
The bottom line is that to get the most out of your data from a business perspective, you need the following:
- An understanding of what you want to achieve and why BI is important as a key support to you and your organization's goals.
- Collaboration between business units and IT to help optimize information access.
- Insights into what you might require in the long run, whether from a business or data perspective.
- A broad understanding of your corporate culture, as different cultures will thrive using different types of BI applications.
- Visibility into your data assets and what you hope to glean from them.
All of these aspects apply to both OSBI and BI implementations. Understanding the value proposition of BI and the key benefits of OS (which will be discussed in the upcoming chapters) can help you decide whether it makes sense to take the plunge, or select a more traditional approach to BI adoption. Let's look at all of this in more detail by breaking down the strategic side of BI deployments -- from a business perspective.
One of the reasons I think it is so important to take a step away from open source BI and look at BI strategy and the general value of BI is because I speak with a lot of organizations about BI, whether conducting industry research, providing advisory services, or putting together case studies. One of the issues that constantly come up is the fact that the industry is difficult to navigate. I am often asked, "Which product should I select?" Or, "I read an article about vendor X, should I implement their solution?" And in most situations, these questions require a lot more knowledge about the current IT infrastructure that exists, business needs, etc.
In addition, some of these same companies are new to BI. In these cases, they have been using Excel and it isn't working for them anymore -- some even call it their "Excel Hell." According to these companies, being able to extract information, put it in a spreadsheet, and manipulate it to your heart's content ends up being the antithesis to developing a strong data infrastructure, or ensuring data validity, privacy, security, and accuracy. Consequently, where spreadsheets have been beneficial for the individual, the corporation is starting to suffer due to a lack of consistency across the organization.
This leads to a broader understanding of why many companies look at BI, which is basically to address the uncertainties that arise due to a lack of trust in the data that is available. However, there are also many other reasons. Let's explore some of the reasons behind BI's increasing importance within organizations and how its use is becoming essential to ensuring strategic decision making.
Goals of consolidating information across disparate data sources
The Excel example just discussed is a good segue into a discussion of the need for data consolidation. There is a common example used by BI practitioners and solution providers to illustrate why Excel on a larger scale does not work and why it is important to develop a secure and centralized access point to business information. Basically, it goes something like this. A planning meeting is called to help identify whether or not the said company is on track mid-year and to identify whether it will meet its overall sales goals. Each department sends a delegate with a prepared presentation to identify sales, accounts receivable, accounts payable, supply chain, partner networks, etc. And each representative has taken the time to prepare the key metrics related to his or her business function by exporting information into Excel, adding calculations based on departmental business rules, and manipulating it to suit individualized purposes.
When all of these decision makers come together and present their information, they find that their numbers don't match up. Basically, it becomes impossible to identify performance when the validity of the data is questionable. Questions arise, such as, which department is right? How can we trust our information? What calculations were applied? And where do we go from here? These are all valid questions that need to be addressed if organizations want to develop a consistent way of managing their data.
There is a constant debate over whether creating a "single view of the truth" is actually possible, which a general Internet search will show you based on the number of results. Although it falls outside the scope of this book, it is still important to know that even though information is consolidated and needs to be accurate and valid, that one version of the truth does not exist. Each department does have its own view of data and business entities. What does exist is the potential to create an accurate, centralized repository that is monitored for validity in a secure way. In this sense, consolidated data provides a better view of performance across the organization, which is not limited to a department or business function.
Better business visibility
All of this leads to better overall visibility. Once information can be trusted and accessed across the organization, it can be used to provide better visibility. Organizations used to silos can now gain a broader understanding of what is happening within their companies. On a high level this sounds simple, but when we look more deeply we see that this is actually a key reason that many organizations consider developing a data warehouse. Some companies actually start their BI projects with the goal of developing a strong data warehouse infrastructure, and the reason is to achieve exactly what we have been discussing -- a centralized access point to information to provide better customer service, product allocation and visibility into broader business operations.
Traditionally, many businesses develop operational systems separately, with information stored separately as well. Therefore, the only way to understand what is occurring beyond HR, finance or supply chain is to create a central data store. Once companies want to understand how finance and accounting intersects with sales and distribution, or how customer satisfaction can be used to up-sell and expand sales, the ability to view information across the organization in a centralized repository becomes essential.
Another reason organizations look at BI is to help mitigate risk. In the past, much risk management within BI remained within the realm of finance, insurance, and banking, but most organizations need to assess potential risk and help mitigate its effects on the organization. Within BI, this goes beyond information visibility and means using predictive modeling and other advanced statistical models to ensure that customers with accounts past due are not allowed to submit new orders unless it is known beforehand, or that insurance claims aren't being submitted fraudulently.
The National Health Care Anti-Fraud Association (NHCAA) estimates that in 2010, 3% of all health care spending -- or $68 billion -- was lost to health care fraud in the United States.This makes fraud detection in health care extremely important, especially when you consider that if you are paying for insurance in the United States, part of your insurance premiums are probably being paid to cover the instances of fraud that occur, making this relevant beyond health care insurance providers. Although risk might not be as apparent within all industries or organizations, many businesses still want to make sure they do not invest their resources in unprofitable products or projects. In these cases, the specific challenges of risk assessment may not be as great but still require a strategic development and use of BI and analytics.