The Principles and Practices to Improve Big Data ROI

The golden rule of analytics is the continuous learning and improvement.

“Big data” refers to the large amounts of information that has become accessible thanks to the Internet. Big Data is useful only if its information content is evaluated for accuracy, relevance, and timeliness. What used to be called “knowledge based enterprises” are designed to transform unevaluated information (raw data) into information whose accuracy and authenticity are verified, and then information is transformed to knowledge, and business insight as well. It is still a matter of manipulating information to make it usable. But with low ROI of Big Data investment in most of organizations, how to improve the way to manage it right?

The golden rule of analytics is the continuous learning and improvement: What needs to change is the desire for quasi-perfection, to solve every missing piece of data, to try and be as perfect as possible when building models. Nowadays tools and techniques exist to change the way we operate, to explore data, discover, and experiment on data driven ideas, quickly drop ideas that fail and exploit those that succeed. The new world of Big Analytics will be where strategic business questions are formulated first, data requirements are defined next, and finally, Analytical solutions are designed and implemented. A golden rule in analytics is the continuous learning and improvement available for all involved in the process from the data collectors to the analysts to the decision makers. Proper feedback mechanisms built into every analytical plan can be as informative as the results of the analysis.
1) Garbage in, garbage out; fighting for clean and the right data never ends
2) Brilliant analytics does not trump bad decision-makers
3) The only thing that is certain is there is no certainty, probably

Analytics is an ongoing business capability: Big Data Analytics shouldn’t be handled as one time project, but a continuous effort for long term. Well, there is entrenched idea that, given a specific purpose for data, we extract the data that is needed, the balance of the effort having no additional future value. There is no learning. This is because the data giving rise to insights is contextually constrained for a specific purpose. The pathetic organizational impulse is to discard the structural and relational bedrock - the part that adds meaning and relevance to the data for purposes outside the original parameters. It's an incredible waste of resources that occurs all the time, just  like building a car to get to one destination; and then building another car to get to the next. The asset isn't the finished data at all, but the foundation that gave rise to it; speaking entirely in terms of data management, some effort is needed to preserve and enhance the value of original capital investment.

Intuition plays an important role in Analytics: Leverage human intuition to look at two or more levels of meaningfulness of data. Big Data can go beyond pattern recognition, recognizing patterns in the data, to noticing the possible relationships among diverse bits of data. One way of doing this is to look at data from two or more levels of meaningfulness. One help in doing this is to let the human's intuition get involved. Computers can help do this if multiple levels of meta-ness can be explored. Such has not been done because computers, designed to do arithmetic, perform very poorly when applied to complex, multi-attribute, multi-relationship networks of meaning.

Big Data deployment shows the overall trends of openness and inclusiveness: Data has the potential to make society more inclusive by giving voice to those that might otherwise be systematically disenfranchised. One way to remove people from a conversation or to limit their involvement is to take away information about them particularly in relation to the consequences of policies and decisions. In order to overcome barriers to sharing, it is necessary to create a valuation scheme for the data such that it carries inherent merit -In our world, people really do exist as data, and our systems are optimized to make life possible through the processing of information. The free-exchange of data represents a democratization of decision-making given that this is precisely what the data powers. The limitations we place over data has significance on the deployment of capital. There are a few interesting examples of some data bucking the overall trend to increased openness (companies attempting to buy crowd-sourced data organizations as an example). These appear to occur where someone has looked at the data and realized that is was significantly more "value-add" than supposed.

"Big Data" needs a paradigm shift to be fully utilized by changing mindsets (from linear to nonlinear, exclusive to inclusive), the way to do things (overcome "We always do things like this" mentality), and keep the end -business objectives in the mind. The overarching purpose of technology should be to empower people to transcend the current level of thinking that created the existing problems in the first place. So the domain logic of Big Data systems should account for this fundamental need rather than merely support status quo thinking. Climbing that great Big Data mountain has nothing to do with getting to the top, but embracing the changes and transformations made to overcome the challenge in building a high-intelligent and high-performing organization.

Posted in: Big Data,KPIs,Principle/Practice