What is analytics?
What’s is the difference between big data, analytics, and business intelligence, advanced analytics, predictive analytics, enterprise information systems, decision support systems?
Answer: everybody has an opinion, but nobody agrees, and you shouldn’t care.
Industry analysts, experts, vendors, and practitioners have created many different definitions of these terms, and how they relate to each other. The result is so much disagreement that detailed definitions are useless.
Instead: ignore the definitions, focus on the business needs of a particular project, and choose the technology that fits best.
There’s always a new term.
After a while, any term associated with analytics starts sounding dated, and people want to come up with a new one. In particular, much tends to be made in the industry of the difference between “backward-looking” old technologies and “forward-looking” new ones such as predictive analytics.
But analytics has always been “forward-looking” — what has changed over time is the sophistication of the technology available to do this.
Concentrate on the actual problem at hand
Business people typically use these terms more or less interchangeably to refer to what they need: better insights in order to run the business better.
If you care about analytics success, forget about the nomenclature wars, and focus on two things: the business goals of the project; and the specific technology to be used.
The business goals for analytics tend to be timeless: greater efficiencies, enhanced customer experiences, new opportunities, greater profits, and increased market share.
The technology that’s available, however, is constantly changing and improving. New terms are invented to help explain what's different from what came before—with a hefty dose of marketing spin thrown in (e.g. data lakes vs data warehouses).
But all these technologies are just different tools, just like spanners, wrenches, and hammers are all useful tools for building a home.
The problem is that many people confuse the technologies available with the business problems to be solved, and end up trying to make home improvements by talking about “hammer problems” or “spanner problems”.
This is a recipe for trouble, because it naturally leads to a focus on the tools rather than the business issue at hand.
To be successful with analytics, separate the business goals of each project from the technologies that may be needed to meet those goals—and you can call it whatever you like!
Global SaaS Marketing Leader | Product Storyteller | AI for Finance
4 年nice piece Timo. I always suggest organizations look in the mirror and ask,"what does advanced analytics mean for me?" Love your stuff.
Accelerated Sales Expert for RISE
4 年"ignore the definitions, focus on the business needs" Yes, it is that simple for the user or client, but not for the implementation responsible. It, analytics, BI, BigData, Innovation or what ever are takes de project need to understand these terms very well in order to go to the market for products and services.
Very good information.
Director of Industry Research @ Industry Knowledge Graph LLC | MBA Visit IndustryKG.com
4 年Thanks, Timo. I enjoyed the article except for the last sentence - hence, I gently offer an alternative for your consideration: "To be successful with analytics, separate the business goals of each project from the technologies - while ensuring the business goals are in sync with overall corporate strategies including the efficient sharing of data (of all types) between business units (such as that provided by knowledge graphs)
Director Industrial Business Unit at Amorph Systems
4 年You state that "analytics has always been “forward-looking”". As one more opinion :-) from my perspective: backward looking and forward looking data strongly differ in their source, in meta data like probability or error, and presumably in their format and size. The backward-source may be an environmental sensor or a digitized client behavior. For the forward-source it may be useless or at least inefficient to create the same data with a predictive sensor or behaviorism. To get exploitable backward-data it is advisable to reduce complexity and separate noise from useful data. Therefore, I do not prefer producing complexity and noise in forward-data. Thus, in my opinion, analytics is always backward-looking, except you analyze forward-data because you do not understand their source or meaning. :-)