Providing the Missing Definition in the Pursuit of Real Insight

FeatureImage.IanPCook.Blog

Seek’s Insight Cloud is solving a big problem in the world of data analytics and decision making. The term “insight” gets used a lot. Like, a lot. The problem, however, is that we never see a clear, technical definition of the term as it pertains to business intelligence and platforms. We take inspiration from the intellectual meaning of the word and apply technical rigor to offer truly useful insights in a software tool. Let’s get into exactly what I mean. 

 

A true, usable business insight must satisfy three crucial criteria: it must be measurable, shareable, and actionable. The Insight Cloud is specifically designed to encompass these characteristics. 

 

Making an insight measurable is essential for capitalizing on the true potential of informed decision-making. First, and perhaps most obviously, a measurable insight relies on available data, rather than hypothetical or anecdotal accounts. This lets the audience of the insights ensure that the resulting metric directly correlates to business outcomes; and enforces consistency and reliability over time. Ideally, an insight should provide a clear explanation of the underlying cause behind a particular phenomenon or trend. A measurable insight goes beyond simply identifying or describing an occurrence and delves into the reasons behind it, offering a deeper understanding of the situation. 

 

Let’s be concrete here. In the world of consumer-packaged goods (CPG) being sold through major retailers, an item being out of stock is a major concern: the store loses a sale, the producer loses revenue and a possible repeat customer if the person picks up a competing product. In fact, it can get even more expensive for the CPG company since the retailer charges them for going out of stock and rushing to fill unplanned outages costs more than the standard item replenishment. Years of experience in the CPG industry could be required to understand not just the implications of an item being out of stock, but the process to understand how to identify when that might happen, and the best process to proactively prevent the problem. Writing down the expert’s understanding of how to identify, ascribe a value to, and what ordering must be done to intervene is what we consider the measurable part of an insight. 

 

Second, measurability demands reproducibility, lest it disappear as a fleeting observation. The best way to reproduce insights gleaned from data is through code. Two of the most common coding languages for extracting meaning from data are SQL and Python ( often a combination of the two). By recording the insight in code, businesses commit to institutional memory the expertise that drove the insight, foster transparency, and ensure that the insight remains quantifiable and reproducible. 

 

This is the first step the Insight Cloud takes to deliver real definition: housing insights as code that can quickly and consistently return the results that drive business outcomes. Data models are created, stored, and versioned to allow you to call on them or update them at any time, right within our platform. 

 

Insights with an audience of one have limited value, which is why shareability is the next necessary piece of our definition. A shareable insight is one that can be easily distributed; and implemented across teams and departments within an organization. It should maintain a consistent definition and calculation method (its measurability) to keep the entire company on the same page. Furthermore, a shareable metric fosters collaboration by providing a common point of reference, encouraging cross-functional teams to work together and make decisions based on a shared understanding. Lastly, a shareable metric should integrate smoothly with a company’s existing tools, facilitating smooth sharing and analysis of the metric among your stakeholders. 

 

Considering the out-of-stock issue from before, sharing the expertise for preventing the costly problem is simple. The Insight Cloud packages up CPG expert’s model of the problem and solution into an application that runs on top of the data a company already uses. Seek’s Insight Cloud bundles the code that generates an insight and deploys it to where you already have your data. Being an Elite Snowflake partner, we are pioneering the use of Native Apps to run code (like SQL and Python) directly within a Snowflake Warehouse, creating the schema, tables, and fields that constitute the core of your insight. This means we can deliver the insights that Seek has generated according to our decades of experience directly onto your data, without any need for further data movement or transformation. Once built, an insight can be distributed to any team with the right data, in a matter of a few clicks. 

 

The final piece is the most important: making an insight actionable. Not only should the results of analysis be clearly presented, they should directly inform decisions that can be made to improve the business. This is where the Insight Cloud moves beyond standard dashboards, as a result of incorporating the prior two pieces of the puzzle. Dashboards grow and die at a dizzying speed, in large part because they are requested by people who have the technical skill, but not the business expertise to make them truly valuable, and are kept siloed from other parts of your organization. In too many cases, a chart or diagram shows a trend but is related to some result that is not within a decision-maker’s ability to control. The trend is thus informative, but not directly actionable. In the case of the Insight Cloud, the business expertise has been captured in the original code or data model, and that expertise has been deployed to any team that needs it, and only then targeted visualizations are layered on. 

 

Relying on our example again, the Insight Cloud generates a visualization of the analysis run to evaluate the size and scope of the CPG company’s potential out of stock issue. Key metrics include the dollars at risk from going out of stock, a list of those products at risk of not being on the shelf, and details about the stores where the problem may occur, among other pieces of information. Notice, however, that these items comprise everything required to generate a list of goods to immediately order and distribute to avoid the loss. The Insight Cloud thus directly communicates the key points of the metrics defined by the expert, scaled to work at any company, without relying on access to a person, in a way that can be immediately acted upon. 

 

The next time a tool, platform, or person promises you insights, ask them to define precisely what that means. We never got a satisfactory answer, so we built it. 

Picture of Ian P. Cook, PhD
Ian P. Cook, PhD
Data-obsessed by nature. Political junkie by training. Experiment-driven by disposition, and experience. All my work involves identifying the right question to ask, using right data and building the right tools to attack it, then crafting the best way to communicate the results. I've succeeded in each of these areas in numerous professional arenas, but never tire of improving.

Leave a Reply

Your email address will not be published. Required fields are marked *