In Part 2 of this blog series, we looked at issues related to data abundance, scarcity, and accessibility, as well as how a properly implemented analytics tools improves an organization’s ability to measure and understand its data, and to make decisions based on said data. In this third and final installment, we will be looking at how the capabilities of analytics can be used together in the data-informed decision-making process.
Capabilities in Concert
Data analytics capabilities often fall into one or more of four specific but often overlapping categories:
Diagnostic: this category examines data or content in order to identify why something happened. It commonly includes techniques such as drill-down ability, data discovery, data mining, and correlations.
Descriptive: this category involves the use of data processing to create a summarization of historical data to provide useful information and, if needed, to prepare said data for additional analysis.
Predictive: this more category makes predictions about future likelihoods—trends, changes, events, etc. It commonly employs various techniques— such as modeling, statistics, data mining, machine learning, and artificial intelligence—to analyze past and present data to identify future outcomes.
Prescriptive: this category is an extension of both descriptive and predictive analytics. It is dedicated to identifying best courses of action for a situation based on the data available.
When the above capability categories are used in concert, they create a cyclical sequence that both drives and improves decision making:
Reports are given on the results/data of past actions and/or events (descriptive)
These reports create awareness; a decision must now be made (descriptive & diagnostic)
The scope and context of the decision is understood (descriptive & diagnostic)
Likely future outcomes are identified (predictive)
The best course of action is chosen (prescriptive)
Reports are created on the results of that course of action (descriptive)
The process repeats
The Right Data for the Job
While the abovementioned capabilities and processes are great, it is important to keep in mind that they are all dependent on one thing—the validity of the data being collected and analyzed. Bad data leads to bad report results… which leads to misinformed decisions that can cost an organization in both time and revenue. In fact, bad (or improperly analyzed) data is estimated to cost industries in the United States a staggering $3 trillion every year.
To ensure that an organization avoids the pitfalls of decisions being made with bad data, it is important to understand how data is being used, how it has been collected, at what level it has been validated, and its potential worth/value:
Prototype: this information has been created by end users, with little to no oversight. Users have the power to integrate different data sources, make their own calculations, and draw their own conclusions. Therefore, it may be questionable and/or unreliable.
Limited Production: this information has also been created by end users in a somewhat controlled environment. It is definitely worth sharing. However, it has not been properly validated.
Published: this information has been gathered in a properly controlled environment and has undergone a rigorous validation process. Therefore, it can be disseminated with confidence as official information.
Best-in-class solutions support all levels of these dynamic analytics environments, including self-service data visualization and data preparation tool sets.
Capabilities in Concert
Thank you for following along with our three-part blog series on how analytics can change every part of your organization. Here at Mashey, we know that success is a journey of increasing the impact of your business and the complexity of your data and analytics. Traveling this path to success means that every member of your organization has access to the reliable information that they need, when they need it, so that everyone (from leadership to sales to support and beyond) can make decisions with confidence.