25 SMART Data - By Dr. Singh

SMART Data - By Dr. Singh

June 06, 2017

We have heard about SMART goals. It seems to me in this world of big data, analytics, cognitive computing and artificial intelligence that we need SMART data as well. George Doran and Robert Rubin have reviewed the acronym and what it could suggest to us, making it easier for us to understand what shared goal-making should be.

What does SMART data mean? The word SMART here too is used as an acronym, each letter signifying certain attributes to the data feeds and reports that we would like to have on a regular basis. Data has meant various things to various people. And if we study the significances that each letter could be attributed to in different minds, it might make it clear that it is like the proverbial elephant to the five wise but blind men who each perceived the creature differently depending on which part of it they touched.

These are the possible significances that come to mind:

  • S - could signify Simple, Sensible, Significant, Snapshot
  • M - could mean Meaningful, Metrics, Measured, Mindful, Mined, Mine-able, Matrix-abled
  • A - could suggest Aligned to goals or interest, Accessible, Available, Actionable, Accurate, Agreed upon, Arrayed, Arranged, Analyzable
  • R - could stand for Relevant, Results-based, Repositoried, Resourced, Related, Relayed, Right, Regular
  • T - could hint Tested, Timely, Time-sensitive, Time/Cost limited, Time-limited, Time-based, Tagged, Testable, Timed

We may add two more letters to the SMART approach to make it SMARTER, the E and R standing for Evaluated and Reviewed or Reported, respectively.

The question then is how to make our data truly SMART or SMARTER? Let us expand on these definitions to see how we can truly drill this down.

  1. Snapshot: The data should give the gestalt in one screen, one picture, whether with the help of graphics or simple figures. Like a physician looking at a patient’s vital signs and immediately figuring how healthy or sick the patient might be. To accomplish this, several questions must be asked:
    • How essential is the data presented to the needs of the recipient?
    • Is the data coherent and uncluttered?
    • Does it assist the receiver or makes him tired and weary while navigating through it?
    • How elegant and sophisticated is the design of the presentation?
    • Is it thought-provoking and thoughtful?
    • Is the approach simple or convoluted, no matter how many indices are computed to arrive at it?
    • Above all, can the individual understand it at one glance?
  2. Metrics: Across the organization, the numbers and computations that are universally understood. Again, going back at the physician analogy, the significance of each vital sign should be the same no matter who the provider is.
    • Does everyone understand these numbers the same way or are there different interpretations?
    • Can they be communicated to everyone so that they appreciate them the same way?
    • Are they critical to the organization?
    • Are these numbers used on a regular basis to determine the health of the organization?
    • Do they align with the goals of the organization and serve as milestones?
  3. Accurate: Garbage in, garbage out. How often have we heard that? Perhaps the most important element of this acronym is the correctness of the data feed into any software or analysis.
    • Is the data precise?
    • Is the data correct and correctable?
    • Can the system detect the errors if any creep in?
    • Can it be analyzed if it is free of errors?
    • Is the data segregated, classified, organized and tagged in a manner that keeps the accuracy and analysis intact in the present and the future?
  4. Relayed: Having the data has no meaning if it is not disseminated immediately to all stakeholders. Once the data is shared, it should be part of the process by which the sharing is done on a regular basis and can be pulled out of the library intact in future as needed. It should be available and accessible always, at any time.
    • Is the data part of a system where data sharing is normal and habitual?
    • Is the sharing of data of significance to the recipient?
    • Is the data shared used or needed by the recipient?
    • Is the data connected to the data feeds of the past and across the various departments?
    • Does the data grow as it is shared in a more meaningful manner?
    • Can it be drilled further for more details?
  5. Timely: Data sharing and accuracy can have no meaning if the data is not actionable. And that can happen only if it is relayed real-time.
    • Can the data influence actions and activities immediately?
    • Is it updated even as the activities are adjusted?
    • Can it be confirmed as free of errors immediately?
    • Can it help the organization pivot readily?
    • Does it help the organization move faster towards its goals?
    • Does it lead to Streaming Analytics and allow Edge Computing?
           

This analysis of SMART data above would be incomplete if it were not SMART in itself. Out of the diverse array of terms that can be used to depict each letter in the acronym, perhaps the most important are Snapshot, Metrics, Accurate, Relayed and Timely.

Good data should be available, accessible and actionable. We should be able to confirm its integrity, disseminate it instantaneously while being able to analyze it in a coherent manner. Good data should follow the standards of organization. Is it individually understood and universally appreciated? Only then will it lead to strong and precise analysis and intelligence that can help save lives and prevent injuries or accidents.