Let your analytics grow together with your business
The era of digitalization has brought analytics to many people’s vocabulary who have not even known about it before. Machines, processes, everything is getting smarter; we have cheap storage space, computation power and analytics to thank for that!
We are bringing intelligence into places and machines where it was earlier considered to be just way too expensive - and enabling at the same time some of our old business cases that were thought to be impossible. Consider the self-driving cars as an example. If the hardware is considered to be the backbone of a machine, then the algorithms really lit it up.
The motivation to run analytics is in business case - naturally, when looking from the business economic perspective. We have a problem that needs to overcome, but that cannot be done without having some more information about the problem, meaning data. Data needs to be stored and collected somewhere in order to build a better and more wholesome picture of the problem. Sometimes this may lead to business models around data itself, as discussed in Joni Lehtonen´s blog post and in this article in the Harvard Business Review. We do not always need to collect huge amounts of data to find a solution with an acceptable probability, but then sometimes it is necessary. This leads to something that is way too often forgotten and we may try to hit a nail with a sledgehammer when a basic hammer would have been sufficient.
Sometimes the analytics touch such a critical process point that failure in obtaining data from the process or information out of that data or feedbacking that information back to the process would lead to huge monetary losses. Similarly, if the analytics only touch a small portion of a bigger picture and what we are actually interested about is a Key Performance Indicator or KPI from that process, it is not always necessary to have all the data sent to cloud nor run all the analytics there. Although it may seem at the moment that most of the analytics solutions are cloud-focused, the intelligence will spread from the cloud all the way to the perimeters of the connected devices - to the fog. Important is that our analytics capabilities can do the same.
To me, being able to scale vertically means also that one can bring in parallel processing capabilities to reduce the time spent on running the algorithms on huge data sets - but only when one really needs it. I like to think analytics platform as a toolbox; one needs to have the correct tool to match the need. It is possible to manage without the correct tool, but it won’t make your life easier.
Start collecting the data, use the proper tool to analyse that and let your analytics grow together with your business. A good toolbox should allow you to have this level of flexibility, and adjust to the terms set by your business, not the other way around.
- - -