Presentation vs. Persistence

People often struggle with where Analytics should fall within the hierarchy of an organization.  When you boil it down to its most basic components it becomes clearer: Analysts themselves are members of the business community, they are part of the strategic fabric of the company.  The main weapon in their arsenal is data.  The data is technically the domain of Information Technology.  This sets up a unique and codependent relationship.

It is safe to say that a modern organization can easily generate and find relevance in more data than ever before.  This acceleration of rich data leads to increased demand.

As digital breadcrumbs continue to drop - such as geospatial tags, personal preferences, and the smorgasbord of the quantified self movement the analytic community develops a hunger that outpaces traditional data persistence strategies.  

Many of us are familiar with the pyramid of data, information, and knowledge.  The base of the pyramid implies a vast volume of lower value data that does not begin to provide its riches until it is contextualized and aggregated as information, and applied as specialized knowledge.

There are many organizational roles to be found as you work through the pyramid, but most can agree that the traditional IT roles of ETL, Data Modeling, and System Architecture are the domain of the base, while analysts (business, financial, clinical, or otherwise) ply their trade near the apex.  

After working in organizations that matrix these roles in numerous ways, it has become very apparent that one of the most successful ways to fuel the information-based organization is to consider two separate but equally important camps: Persistence and Presentation.   

Persistence allows the technology department to focus on moving, cleansing, and storing quality data in a manner that doesn't confound the analyst community.  A world class persistence team can respond quickly and thoroughly to requests for data, accept feedback from consumers, and make it available to the presentation teams.  The Presentation teams can then spend less time focusing on data acquisition, creation, and quality - because there is trust that anything that has been persisted is accurate.  

The blurring of these two concepts has led to failed initiatives across many different organizations.  This blurring occurs when the analysts (Presentation) try to hard to service their own data needs creating silos of ungoverned data.  Access databases, business critical spreadsheets, and data assets tied to a single employee name are all warning signs.  Another way this blurring occurs is when IT roles (Persistence) attempt to outsmart the analytical community.  Concern should be raised when members of the technology team argue that "the business" doesn't need a particular kind of data, or challenge the priority of an analytical request after it has been approved by leadership.  

A healthy information-based organization recognizes that there are inherent checks and balances between Persistence and Presentation.

The resulting apparent dichotomy need not be frightening to leadership.  It is this separate but equal organizational distribution that leads to healthy discourse, and optimized processes.  A correlation can be made to the United States Government.  It can be argued that when a single party is in control of both the executive and legislative branch, much appears to be accomplished.  If you are a member of the opposing party it seems as if nothing is getting done - and it is often overturned in the inevitable snap-back of the following election cycle.  If either the Presentation or Persistence team operates unilaterally, the resulting lack of checks and balances results in a lot getting done, but much of it does not end in lasting strategic improvement.

In conclusion, a strong case can be made that technology professionals who are focused on moving and persisting data should be fully enabled members of an Information Technology department, while analytical teams (including Business Intelligence Analysts) should be firmly embedded within the business as they are the providers of the context and application that moves that data to knowledge.