top of page
strakanakis

Too Much Data, So Little Insight and the Role of Analytics

If I had to take two Points as given, even if this is against the core belief of any decent Analyst, it would be the following:

Too much Information is not Information.

Data is far from Insight.


Complicated Mechanics
In Dynamic Complicated Systems even the Smallest Parts can have Great Impact

Actually this is what we struggle with in our Time, too much “Information”. Paraphrasing John Naisbitt and taking his quote one step back I would state that “We drown on Data but Starve for Insight”. As the muddy waters of a dirty Californian river were hiding the Golden Nuggets from the Gold-Seekers in the Middle of 19th Century, the abundance of Data is hiding the Golden Nugget of our Times, which is none other than information. This is where Analytics step-in and do their part, filtering the muddy river of Data and clearing out the dirt in order to retrieve the golden nuggets of Actionable Insights.

I want to be clear that what I say is not meant to underestimate the Value of Data, but to stress out that Data is not enough. It is beyond any doubt that Data is required for Analytics. Still the Presentation of Data, in the form of long tables crowded with Metrics that are as much related as the people squeezed into the London Underground on a Monday morning, is far from being considered Analytics. One of the main deficiencies of Data is that they are static, not connected and lack context.

At this point I need to clarify that Data is considered anything in a Raw form, that is not processed by Analysis nor Imported to a Model.

Data is to Analytics what Marble is to the Parthenon. While the Parthenon would never exist without the marble, on the other hand the same marble that was vital for the building would never stand out of a pile of rocks if it was not worked in order to stand as a physical representation of a mental model. Because this is the beauty of Parthenon. That it represents Mathematical Concepts that capture the cornerstone of the Ancient Greek Civilization.

In order for Data to be shaped to Analytics it needs to be refined from a model that proceeds the Actual Analysis of the Numbers. This is what we employ in the form of Risk Analysis or what the Stoics where stating as “Premeditatio Mallorum”. Before any Analysis can be employed a Mental model of all involved Risks and Targeted Objectives are clearly defined. First we need to consider the underlying Risks that are built into the our pursue our Objectives. Then using them as our Building Blocks we need to develop a Model, a Scenario if you like. The next step would be to identify the Key Points of the Business that would be affected and the Metrics that would capture this effect and how they are expected to deviate

 After all of the above are done it can be Quantified, within a range, the possible Effect of each decision and Thresholds can be set that would be separating the desirable Outcome from the negative. This is vital to be done, as even if it can not Guarantee Success it can Safeguard against Fatal Failure. This becomes more and more challenging due to the increased Complexity of the systems we operate in, or this is what we tend to believe.

The real problem does not derive from the increased complexity. It rather derives from our attitude to that and this can be clear after we consider the following. The increased complexity of the systems Analyzed, as well as in the Data Availability, has pushed us into asking more Questions, which is still positive. The real problem is that we tend to React to this by giving more Answers. Even if it sounds contradicting, when the Questions multiply the objective is to give even less answers, but more complete answers. More connected.

Our failure to deal with the increased answers results in the extended List of KPI’s and Noisy Dashboards. This happens before we try to answer all Questions at once and isolated. 

What we should be doing instead is to try to understand the interactions and co-dependencies among KPI’s and what is causing their relative movement.

KPI’s were never supposed to be the answer they are supposed to be working as Alerts and as the compass to develop educated Hypothesis that is Tested through targeted testing rather than blindly looking in the dark.

Since my early days as an Analyst I felt the need to have a more efficient high level approach that will be acting as the first line of Analysis, Identifying trends in the making, rather than tracking Trends that have been already established.

This need has driven me to Develop the Key Revenue Drivers (KRD) Approach. KRD consist a set of KPI’s, that their inter-relation has been mapped and their relative movement captures the fundamental mechanics of the system Analysed. Focusing on the Mechanics rather than the End Result of a System gives great value. This is because the End Result consists the Snapshot at a point in Time, while the understanding of the Mechanics focuses on the Continuity of Performance.

While the Snapshot can depict where we Stand, the Mechanics can project also where we are Heading to. The Analysis of their Relative Movement serves as a first class indicator on underlying issues or opportunities, even before being fully depicted in the Customer Base Performance. This is vital because if we wait to act based on established trends, we will be constantly running after departed trains.

More details for the KRD I developed and apply in Gaming for over 4 Years, you can find in my article

To summarize all of the above I will refer to the following quote by Clement Bernard:

“The experimenter that does not know what he is looking for, will not understand what he finds.”



2 views0 comments

Comments


bottom of page