- WE’RE LIVING IN A METRICS-DRIVEN WORLD
We live in a data-driven world. For Fortune 500 companies, the value of data is clear and compelling. They invest millions of dollars annually in information systems that improve their performance and outcomes. Independent businesses have the same need to be data-driven; however there’s a persistent entrepreneurial resistance to becoming truly metrics-driven. Founders are often tempted to postpone building necessary metrics in favor of spending time and resources on building products.
While that might work in the short-term, it will very quickly come back to haunt them. Very few companies have successfully achieved exponential growth, raised capital, or negotiated strong exits without first having a solid analytics model that has been iterated upon for many months or years.
“The Analytics-Driven Organization”
As companies become analytics-driven, they aren’t just enjoying incremental improvements. The benefits enabled by analytical data processing becomes the heart of the business – enabling new applications and business processes, using a variety of data sources and analytical solutions – giving insight into their data never dreamed of and giving them a great competitive advantage.
Below you’ll find a description of some of the types of data analytic insight types, and common use cases for each.
Descriptive: Descriptive Analytics uses business intelligence and data mining to ask “What has happened?” Descriptive Analytics mines data to provide trending information on past or current events that can give businesses the context they need for future actions. Descriptive Analytics are characterized by the use of KPIs. It drills down into data to uncover details such as the frequency of events, the cost of operations and the root cause of failures. Most traditional business intelligence reporting falls into this realm, but complex and sophisticated analytic techniques also fall into this realm when their purpose is to describe or characterize past events and states. Summary statistics, clustering techniques, and association rules used in market basket analysis are all examples of Descriptive Analytics.
Diagnostic: Diagnostic Analytics examines data or content to answer the question “Why did it happen?” It’s characterized by techniques such as drill-down, data discovery, data mining and correlations. You can think of it as the casual inference and the comparative effect of different variables on a particular outcome. While Descriptive Analytics might be concerned with describing how large or significant a particular outcome is, it’s more focused on determining what factors and events contributed to the outcomes. As more and more cases are included in a particular analysis and more factors or dimensions are included, it may be impossible to determine precise, limited statements regarding sequences and outcomes. Contradictory cases, data sparseness, missing factors (“unknown unknowns”), and data sampling and preparation techniques all contribute to uncertainty and the need to qualify conclusions in Diagnostic Analytics as occurring in a “probability space”. Training algorithms for classification and regression techniques can be seen as falling into this space since they combine the analysis of past events and states with probability distributions. Other examples of Diagnostic Analytics include attribute importance, principle component analysis, sensitivity analysis and conjoint analysis.
Discovery: Discovery Analytics doesn’t begin with a pre-definition but rather with a goal. It approaches the data in an iterative process of “explore, discover, verify and operationalize.” This method uncovers new insights and then builds and operationalizes new analytic models that provide value back to the business. The key to delivering the most value through Discovery Analytics is to enable as many users as possible across the organization to participate in it to harness the collective intelligence. Discovery Analytics searches for patterns or specific items in a data set. It uses applications such as geographical maps, pivot tables and heat maps to make the process of finding patterns or specific items rapid and intuitive. Examples of Discovery Analytics include using advanced analytical geospatial mapping to find location intelligence or frequency analysis to find concentrations of insurance claims to detect fraud.
Predictive: Predictive Analytics asks “What could happen?” It’s used to make predictions about unknown future events. It uses many techniques from data mining, machine learning and artificial intelligence. This type of analytics is all about understanding predictions based on quantitative analysis on data sets. It’s in the realm of “predictive modeling” and statistical evaluation of those models. Examples of Predictive Analytics includes classification models, regression models, Monte Carlo analysis, random forest models and Bayesian analysis. It helps businesses anticipate likely scenarios so they can plan ahead, rather than reacting to what already happened.
Prescriptive: Prescriptive Analytics uses optimization and simulation to ask “What should we do?” It explores a set of possible actions and suggests actions based on Descriptive and Predictive Analyses of complex data. It’s all about automating future actions or decisions which are defined programmatically through an analytical process. The emphasis is on defined future responses or actions and rules that specify what actions to take. While simple threshold based “if then” statements are included in Prescriptive Analytics, highly sophisticated algorithms such as neural nets are also typically in the realm of Prescriptive Analytics because they’re focused on making a specific prediction. Examples include recommendation engines, next best offer analysis, cueing analysis with automated assignment systems and most operations research optimization analyses.
Previous post can be found here.
Next post can be found here.