Leaders manage; they plan, organise, direct and control. Analytics used to be just "looking in  the rear-view mirror".

Two problems with that:-

  1. Not looking where you are going- you might crash!
  2. History may not predicate the future- remember "Black Swans"?

Avoiding a crash

Leaders throughout an organisation need predictive as well as historic analytics. That can vary from simple time-series analysis and regression analysis to complex modelling. At the very least users need automatic alerts to anticipate potential problems and opportunities

This does not just mean data scientists deploying algorithms, machine learning etc. It must involves operational people near to the front-line who can add intuition to analysis and predictions. This is why self-service BI and analytics is a vital component of all enterprise decision making,

Black Swans and decision making

When is an outlier an unrepresentative outlier and when is it potentially the extreme impact of certain kinds of rare and unpredictable events?  Nassim taleb's Black Swan Theory

It is often the person with detailed operational experience and intuition that can tell the difference. The police officer on the beat who against the tide spots the incongruous. The one who does not follow the lemming instinct to run with the herd and fall over the cliff. 

Only 10% to 20% of data in an organization is typically used to inform and provide insights. That is the highly structured, cleansed and reliable data. 

80% to 90% of data is unstructured:-

  • emails
  • letters
  • free-form comment fields
  • conversations
  • Photos
  • Videos

Unless these data are ingested and analysed decision-making will be poor, uninformed and incomplete. It is not necessarily a matter of "big data"- more a case of "complete data".

Complete Data for complete decisions

We have not even considered external data yet- this can be five times the amount of structured and unstructured data in an organisation. It may be 10 times the amount.

Take insurance and claims.

  • Physical location of a solicitor in a claims report- Google street view will show nearby body shops, scrap dealers, second-hand car dealer- a sign of fraud if there were ever one!
  • Time of accident- many infants claimed to be in an accident are during school times. Fraudsters know that a child named in accident form increases chance of accepting whiplash claims

Checking external data can often improve decision making

Analytics for better decision making and execution

  1. Can you analyse past, current data and predict future?
  2. Can you analyse structured, unstructured, internal & external data i.e. complete data?
  3. Can you join multiple and disparate data sources in same reporting and dashboards?
  4. Can you deliver self-service analytics to all personas across an organisation from one BI platform? i.e. -
  5. Information Consumers?- e.g. CEO, line worker, nurse
  6. Information Creators? e.g. HR Manager, production manager, hospital administrator
  7. Analysts? e.g. marketing operations & research, finance analysts,
  8.  Can you apply enterprise strength security  across all users and LOB structures?
  9. Can you embed analytics into enterprise applications and workflows so users have the insights in the context of current key tasks and activities?
  10. Can you both present insight to users and empower them to update data- i.e, apply write-back for audit trails, updating data and initiating action?

Best you can answer yes to 80% of these questions and better that you achieve 100%