Successful companies know that achieving a competitive advantage in today’s market is largely a function of deploying better and more advanced analytics of a growing variety of data. Analytics allows companies to gain insight into the behaviors of consumers and competitors; to understand demand and supply trends; and to evaluate operational performance and a number of other business critical data sets. The technology also helps employees to make fact-based decisions on innovation, marketing, pricing, discounts, logistics, services, and a number of other areas.
The expansion of analytics is also driven by systematic, fully automated data collection and the capture of behavioral data from multiple touch points. Behavioral data provides more evidence of what people do and why they do it than self-reported data, which is typically collected via surveys and focus groups. Behavioral data allows us to see the actual consumer choices and behaviors.
Nearly every interaction today leaves a breadcrumb of information in various systems and formats. As a result, it’s becoming increasingly important to build a conceptual framework that enables enterprises to piece these breadcrumbs together into coherent and comprehensive stories. See below for an example of such an analytics framework.
Different Data Types
Analytics is frequently characterized in terms of breadth (shown along the X axis). This indicates the variety of data types that can be analyzed. Today’s enterprises collect both structured and unstructured data and those two source types have to converge into one analytic system. More often than not, they are treated as fundamentally different and the management and analysis is organized within different groups and departments, thus missing an opportunity to link the insights gathered from both.
Of course the variety within each data type is enormous, but the idea is that unstructured data cannot be treated as a non-standard type of data outside of the business intelligence (BI) and analytics system. It cannot be ignored by IT or business analysts because it is outside of the traditional data warehouse. It has to be an integral part of the analytic strategy and process.
The depth of analytics (shown along the Y axis) is the dimension that characterizes the increasing complexity of analytic techniques and methods that can be applied to the same data types. For example, the same data set can be analyzed both descriptively and predictively and the insights will differ greatly. The descriptive analysis provides insight into what is happening, while the predictive analysis sheds light on what is likely to happen. The depth has to be correlated with the skills and talent available in the enterprise. As more advanced techniques are used, the value extracted from the collected data increases exponentially.
Naturally, both depth and breadth are increasing over time as enterprises continuously learn from working with the data and seek to expand their knowledge. The value of the present conceptual framework is to allow BI and analytic professionals to assess where their organizations stand with respect to those two dimensions and formulate both near and long term strategies that map to the full scope of analytics.
You may ask why I choose the bell curve? As the chart shows, the higher levels of analytics stack on top of the lower levels of analytics. Reporting is at the base, as it is the starting point for all business analytics in every enterprise. It also requires the least advanced skills for information consumption (if done well) and thus is the foundation for pervasive distribution of data for decision-making. Predictive analytics and optimization are at the top as they are highly specialized methods that must be performed by trained statisticians and mathematicians.
However, I want to convey an important message here: if predictive analytics is embedded into the operational reporting, the average intelligence of the entire enterprise will increase. Conversely, if analytics is implemented by statisticians – and the findings are distributed in memos and presentations – the project will impact the knowledge of the few rather than the actions of the many.
So the key to raising the analytic capabilities of the entire enterprise is to embed the more advanced methods into operational reports and dashboards so that all operational employees are empowered to make better fact-based decisions.