Cognitive Analytics
A new dimension for the modern business
Experiences have taught me that people often have varying views and perceptions of the same reality, and this phenomenon occurs universally. Professionals grasp or develop different understandings of the same concept because their perspectives and contexts differ. This results in misalignments, misunderstandings, and inaccurate decision-making.
Insights from data unlock opportunities but only when there is a common understanding of technical jargon and business language.
Which Analytics framework is the common language?
Over the past few years we have learned about descriptive, predictive, and prescriptive analysis as the main stages in the universal analytics framework. They serve as a benchmark to gauge maturity in data-driven capabilities:
- Descriptive Analytics: Reflects on past information to provide insights.
- Diagnostic Analytics: Monitors current data for oversight.
- Predictive Analytics: Leverages data to forecast outcomes.
- Prescriptive Analytics: Extracts insights from data to proactively pursue optimal business outcomes.
A significant gap remains: Traditional analytics don’t guide users toward the ‘right sight’ from datasets before diving into the analysis phase.
Delving deeper into the limits of established analytics “dimensions”
So what’s missing? Well, virtually all employees have acquired increasing experience in BI and analytics practices and software tools. While many are proficient, a common pattern emerges: when approaching a new analysis, users tend to subjectively expect certain types of patterns, insights, and answers based on past results and events. Preconceptions and biased expectations can skew the analysis, leading to biased and incomplete answers.
In business, repeatedly asking the same questions and expecting similar answers is a limiting mindset, especially if growth is the end goal.
Conversations with Kirk Borne and the Graphext Team confirmed my observations. Analysts and end-users of analytics platforms tend to mistakenly approach the Exploratory Data Analysis (EDA) phase anticipating the majority of the answers they seek.
In business, repeatedly asking the same questions and expecting similar answers seems like a limiting mindset, especially if growth is the end goal. Increasingly, the goal should be to equip the end user with tools and methodologies that address this bias in a scenario. A new analytics dimension caters to this evolving mainstream: Cognitive Analytics.
Cognitive Analytics
It focuses on determining the right questions to ask to your data or the right subject or event, within the correct business context. It mimics human curiosity, uncovering valuable signals and facilitating the discovery of fresh insights using machine learning. Cognitive analytics can be thought of as an analytics product with human-like powers and the most effective approach marries algorithms and natural human perception for efficient data exploration.
Two real life business cases where Cognitive Analytics makes sense:
- Retail Customer Segmentation: A common approach for a large retail chain using descriptive and predictive analytics is to segment customers based on purchase history, demographic data, and online behavior. Over time, analysts develop certain preconceptions, e.g., “customers aged 18–25 prefer online shopping over in-store visits.” Such preconceived patterns inaccurately influence future analyses and marketing campaigns. Value: Using Cognitive Analytics the retail chain can ask more nuanced questions about customers. Instead of relying on established segments, the system can explore patterns outside the regular scope. For instance, it might detect that customers aged 18–25 who live near the store prefer to visit if there’s an experiential or event-based promotion. This insight, which goes against established beliefs, leads to effective in-store promotional strategies for this demographic.
- Manufacturing Quality Assurance: The traditional method that a manufacturing firm uses to monitor its assembly line is with IoT and sensors. Their analytics system flags parts and processes that deviate from predefined standards. Over the years, technicians develop a bias that certain machines are “troublemakers” based on historical data. As a result, newer ones might get less scrutiny, potentially overlooking emerging issues. Value: By introducing Cognitive Analytics the firm can assess each machine’s performance without past biases. The system prompts quality assurance teams to ask questions like, “Are there environmental factors affecting machine performance?” or “Are certain issues arising in machines across different age groups?” Instead of just flagging deviations, Cognitive Analytics enables users to find the right questions to ask about why these deviations might be occurring. The firm discovers that certain issues are more related to shifts in ambient factory conditions than the machines’ ages. Addressing these environmental factors at the end improves overall product quality.
In both examples, Cognitive Analytics aids in sidestepping biases that could blind users to valuable insights.
Anyone in Business Intelligence, Data Science, and related fields can harness the potential of Cognitive Analytics. It can bestow both strategic and tactical advantages to organisations by encouraging users to find broader questions to ask. Regardless of whether a company relies on on-premises storage or cloud solutions, this new dimension will redefine the way information is analysed, uncovering novel strategies to address problems more effectively, and paving the way for deeper insights and more streamlined operations.
In a nutshell
The challenge isn’t merely about adopting innovations through software purchases and Cloud/AI consultancy services anymore.
Nor is it about connecting or integrating sources, transforming, processing, visualising, or analysing data.
The challenge now is considering the nuances of human perspectives, under the growing complexities of the digital business landscape.
After extensively applying the primary stages of traditional analytics, now the common pitfall that businesses fall into is not addressing biases and preconceptions. Cognitive Analytics focuses on removing biases by prompting users to ask the right questions rather than jumping to conclusions based on previous patterns or results. It combines human-like curiosity with machine learning to discover fresh insights.
Though cognitive analytics is still in its infancy, it offers a robust solution for uncovering unexpected answers. It represents a paradigm shift from traditional analytics
Grasping the distinctions among these analytics categories and adopting both Cognitive Analytics practices is paramount for a modern thriving business.
