Caveat: the term "single source of truth" has been hijacked by salespeople, consultants, and sales engineers in every industry that relies on or deals in Big Data. While there is a legitimate need for companies to implement the strategy to streamline information sharing, most advice from 3rd parties ends with acknowledging that it is a need with no concrete plan to implement a strategy.
The idea of a Single Source of Truth, or SSOT for short, came about with the advent of sophisticated ERP's and IoT paradigm shift in data collection. With multiple sources of incoming and outgoing data, it became imperative for strategic decision makers at all levels to be well-versed in how to use data to justify business decisions. Some industries are further along than others and gaps exist below the company level in maturity, but there is an understood need that intuition should be backed up by hard data and statistical analysis.
In my profession of mining, Short Interval Control and data historians are used to provide an optimum output. Traditionally in more mature environments, data historians populate SQL databases, either local or cloud-based; this SQL data is then pulled into a data production "front-end" platform like Tableau or Microsoft Power Bi. Once in the data production platform, data can be refreshed and manipulated to give the end user nearly any statistic or Key Performance Indicator they desire.
Where the concept of SSOT comes into play is creating a research tool that combines all available company data into a single workbook with datasets being related through common columns. These relationships are displayed through an Entity-Relationship Diagram, a map of connected tables and relationships. I have attached a sample at the bottom.
Where SSOT shines is as a research tool. Currently, my tool pulls in accounting, supply chain, maintenance ERP, and condition monitoring data into a single workbook that shares common filters and is refreshed daily. Rather than having to manually pull various data sources, standardize and clean data, and perform analysis, these tasks are automated through Power BI for easy retrieval and manipulation through the DAX programming language.
The challenge then becomes one of culture change shifting from data being pushed to end users vs. data being pulled by end users. The latter is a hallmark of a mature company and requires extensive support in the short term to increase the baseline skill level of the end users. Stay tuned for my next blog post on how I help to support this culture change.
As always, if you found this interesting, please share it, and feel free to email any questions you have concerning learned best practices.
Comments