THANK YOU FOR SUBSCRIBING
A Brief Definition of Business Intelligence
By Kudret Soyturk, Head of Business Data Intelligence, Atos [EPA: ATO]
Although the rapid and versatile development of tools has stressed the science side of the process, human aspects such as message selection, presentation of options like colors or icons reflect the design side. This combination of science and design skills makes BI attractive to creative employees, as well as adding value to the business.
Processing more and more data
If we start with one single unique data point, it is not enough to create any intelligence; at least two data points are necessary to reach a conclusion. For example, a single data point indicating that I have two apples is not enough to provide any useful or actionable information, but when I have the second data point everything changes:
- my average daily consumption is one apple, so plan to go to market in two days
- my fridge capacity is three apples so available capacity is one apple, therefore don’t buy more than one apple at the market
As the simple example above shows, the more data we have, the more intelligence we can generate. Driven with this motivation, BI experts are striving to gather as much data as possible, ideally providing the highest level of granularity.
How does data matter?
Until recent decades, data had predominantly been generated internally through manual effort and has come with a significant associated cost. The introduction of recent technologies and sources such as the Internet of Things (IoT), Web click tracking, and Open Data sources have simplified the data collection process and satisfied the need for massive volume. Now it is possible to collect real-time data from numerous devices, internal and external websites, curated libraries of government, market and generic data in addition to the traditional line of business applications. As an example, World Bank Open Data provides 3,000 datasets and 14,000 indicators encompassing microdata, time series statistics, and geospatial data.
BI has the potential to say a lot, but leadership may not be ready to absorb all this information at once
Is having data enough?
Having copious data in many different formats such as database, semi-structured, voice, video can cause some unforeseen problems. When you have a huge volume of data, you need to have effective storage management (online, nearline, offline, etc.) and massive processing power for the exponentially growing calculation load. Recently, those needs have started to be addressed with cloud solutions and provision of those assets as XaaS (Everything as a service, let X be Infrastructure, computing platform, etc.) which optimize the cost of ownership and allows for virtually unlimited capacity.
Massive quantity and variety of data can lead to an unlimited number of hypotheses to test and can even allow you to process data without having any hypothesis in place to discover new intelligence on the fly. Here we start to see the increasing role of high-performance computers, rather than only human capabilities. RPA (robotization and process automation) is to replace humans for most of the rule-based, well-defined activities and perform them at an extreme speed and quality. When the rules are not well-defined and/or unclear, AI (Artificial Intelligence) and ML (Machine Learning) solutions are there to help.
The human side of BI
Although BI touches every technological improvement and is impacted by them, it is a journey with many non-technical aspects. BI can support the highly critical decisions used by leadership. However, what if some assumptions are not correct, the BI organization loses credibility, which is the most important driver for sustainability and the future growth of the BI organization. Maximum effort should be spent in building and sustaining credibility; this effort may include, but not be limited to, test and assurance activities, finely balancing between quick wins and large-scale programs, and agile versus traditional project management.
BI has the potential to say a lot, but leadership may not be ready to absorb all this information at once. A good communication strategy should be established, ideally, there should be an understandable story behind each figure. In the end, people will be more prepared to accept information that has been communicated effectively.
Finally, it is important to prevent overloading with content or messages. When you are rich with data and tools, it is a common tendency to calculate extensive numbers of metrics and try to present all of them. This approach has some disadvantages which jeopardize (instead of support), such as clutter, confusion, and information overload. Keeping a digestible number of metrics, graphs, colors, lines, and words should be the goal.
Some final words
To some extent, everyone who has access to a spreadsheet application can be a BI expert – we need to consider this fact to our benefit and use it as leverage to define and re-define what is a BI Technologist. BI, requiring both technical and non-technical skills, will be a growing interest area to attract distinguished talent, even though the required human activities and practices will change significantly over time. Maybe this everchanging dynamic nature of the human component will be the lure of the business intelligence profession.
Tools That Allow Us To Turn Data Into Business Intelligence
Philip Garland, CIO, PricewaterhouseCoopers
Exceeding Corporate Goals with Business Intelligence
Daniel M Horton, CIO, Sallyport Global
Managing a Major System Change to Reap Organizational and Business Rewards that Extend beyond Technology
Christopher Dowler, CIO, IAT Insurance Group
The Right Conditions for Data Science to Thrive
Andrew McMurtrie, Head of Data, Direct Line Group