We live in the time of the greatest opportunities. Almost every tool and possibility become affordable, and if some decades ago such solutions as BI were privilege now they are the need. Their implementation is a great investment in the future. That is why today we highlight you everything about BI implementation process. So let's start.

What is BI about? 

Business Intelligence is a set of IT technologies for collecting, storing and analyzing data. It allows to provide users with reliable analytics in a convenient format, based on which you can make effective solutions for managing the company's business processes.

All user levels, from employees to founders, have flexible access to the management reporting they need, without IT specialists help.

There are following major BI platforms:

• ETL tools: programs that allow you to load data into DWH from various accounting systems.
• DWH Storage: a complete SQL database for preparing and storing data for analytics.
• OLAP-cubes: a technology that allows you to make any reports in real time (1-5 seconds) and carry out a complete data analysis.
• Client applications: as a rule, users deal with Microsoft Excel PivotTables connected to OLAP cubes for detailed data analysis and building dynamic reports. For surface analysis and visualization of key indicators, WEB-applications are also used, which should support access to reports from any device: computer, tablet, phone.

Why implement business intelligence? 

To start implementation of every new technology and system it is important to understand clearly whether you need it at all and why. To make it easier you should consider all the benefits it can bring to you and problems that can occur. Let’s start with a pleasant thing and discuss all the advantages you get implementing BI:

Before starting: FAQ 

One more important step towards successful BI implementation is to think about concern questions. And, what is no less important, to get answer to them. That is the right think we are going to do now.

We gathered top 5 FAQ regarding BI and it’s implementation. So here we go:

What are the BI implementation threats? 

As we all know, together with all the benefits from every technology come the challenges of its implementation. So it is better to face and solve them in advance. Here are the core difficulties companies meet implementing BI solution:

• High cost: such complex solution are usually not so cheap. Besides the product or subscription cost itself it requires budget for its maintenance, installation, employee trainings after. So, such products are more oriented for large companies and are not always needed and affordable for small ones. However, nowadays there are dozens solutions on the market that will suit all the companies categories with any employee number and budget.

• Problems related time and cost estimation: the second problem is rather spread and in the same time the most solvable. The key of success here is an appropriate planning and reasonable assessment of opportunities.

• Absence of training: to better understand a new system and debug the work of departments the trainings are real must. Lack of them or absence at all can cause serious problems in the future.

• Non-acceptance by the staff: this problem mostly comes from previous one. You employees can have lack of knowledge of how to use a new tool or have lack of motivation to do it. So here is another sign, the training and work with a staff are the must.

Who are the BI tools primary users? 

There are four types of users of business intelligence systems:

• Analysts: the core users. They are few in number and present the company’s brain.
• Top managers: the other important users are top managers of the company. As a rule, they understand that a business analyst can resolve the issue of business health, and therefore they treat BI tools with a certain respect.
• Middle managers: the third type of users are middle managers. The focus of their activities involves solving a narrow range of tasks related to the area of work performed.
• Ordinary employees: there are also employees who need a business analyst to access relevant information and create reports.

What is BI implementation strategy and how to build it?

BI strategy includes all the stages, efforts, decisions you go through implementing BI solution in your company. It consist on the immersion in the BI process, identifying stakeholders and key figures, assessing the situation, setting goals and determining performance indicators that will help you measure your efforts to achieve these goals.

The team leaders define vision, all the organizational processes, architecture, cases and on its base build a roadmap. 

How long does implementation process take? 

It mostly depends on the following factors:

• Company’s size
• Project scope
• Goals

So considering all these facts, implementation may take from few weeks to several month.

What are the key indicator of successful BI implementation? 

The way to successful implementation starts even before its direct beginning. First of all, you can identify KPI, which will help to estimate implementation’s results. The main indicators of success here are:

• Utility - the system performs the function
• Reliability - the system works without failures
• Efficiency - the system works quickly and efficiently
• Expectations - the system meets user expectations

Building BI implementation roadmap 

And now we are finally get to the main challenge: BI implementation roadmap building and the start of implementation itself. Let’s overview these processes on the Ranet OLAP Implementation example. Ranet OLAP is a ready-to-use data analysis tool that helps to create interactive reports, forecasting and business planning. The tool is easily extendable and can be integrated into existed system used by organization. 

Take patience and go on a thorny but fascinating path to BI opportunities.

Step 1: Determination and analysis of requirements 

The first thing to do is to determine requirements. At this stage, it is right to be in contact with a specialist or contractor with market expertise and knowledge of how the industry approaches similar tasks.

The correct method here will be to go from top to bottom - if you automate existing reporting, moving from lower-level specialists, managers and analysts to the top management, then in the end may occur the work was useless, because top managers needed different numbers. Moving from top to bottom, we get the right overview: the CFO knows what he should see, then his request adapts to the level of regional and local managers, and they, in turn, clearly understand what numbers are needed at their level.

Step 2: Data organization 

There are two ways: from general business requirements or from the needs of each unit. In the first case, you must first analyze all business requirements, then work out the needs of each department.

The second approach is iterative - we divide the entire scope of work into separate areas, and describe in detail how analytics and reports will look like for the marketing department, then for finance, HR, and then go through iterations across all departments.

Step 3: The choice of technology stack 

At this stage, it is necessary to identify data sources and clarify whether they contain the necessary information and indicators. Very often it is necessary to refine accounting systems so that the indicators start. When the source pool is assembled, you can navigate to accounting systems, web resources, and company internal systems in order to component-wisely design the architecture and prescribe the role of sources for data transformation.

Any information in the BI-system comes in raw form, and at this stage it depends only on us how accurate and easy-to-read data managers will get at the output.

Step 4: Interface Design 

Employees who use the system appreciate the user-friendly and pleasant-looking interface as deeply as the possibilities that the solution gives. Therefore, projects often introduce a prototyping stage, when we draw the interface forms.

This stage helps the client to understand which graphs are better to use for visualizing certain indicators, which colors to choose, how convenient it is to arrange a filter, etc. After the data transformation stage, this prototype will be sufficient to fill. Otherwise, it fully meets the expectations of business users.

Step 5: System testing 

Here you need to develop test scripts. Take unloading in one of the directions for a given period and accuracy of information on the same data slice from the same accounting system.

Another case is the implementation of the system to a dynamically changing data source, or when we implement a solution to Excel data, but the data loading step needs to be transferred to a newly deployed source, in which everything from the storage structure to the information itself could change. Here, implementation and testing will follow different rules.

Step 6: Team training 

The most important step in preparing to launch a new system is employee training. You need to make training for financiers, marketers, IT specialists and managers. The training includes learning the platform, the possibilities of refinement and solution management, teaching managers to make the most of all the features of the program. To help administrators and users, accompanying documentation is developed: the basic Administrator's Guide and User's Guide, and often training videos.

The most detailed and complex, but useful material is what is usually called Techproject or Report specification. It describes the entire process of moving data from sources to final reporting forms. Do not neglect this document. With it, any newcomer to the team will be able to figure out how the data fall into the first loading layer, and where they are in the output reporting forms. With this material, any change or request for finalizing the system will take a minimum amount of time.

To cut a long story short 

It is challenging to create any informational system. Designing analytical solutions is hampered by the capricious and difficult to work element - data. An experienced team will solve this problem faster without incident. Regardless of whether you prefer single descents on the thresholds of the Amazon or controlled by an instructor, pay the most attention to the actions with the data. Then technical and methodological difficulties will pose a lesser threat, and the future system will be able to solve complex analytical problems without errors.