Turning Insights into Actionable Outcomes
Examine the image below. What do you think it depicts?
If you believe the image depicts a tasty piece of chocolate, you're mistaken.
Displayed here is compound chocolate, a blend of cocoa, vegetable fat, and sweeteners. It's an economical alternative to authentic chocolate, due to its less expensive ingredients. You can find it in budget-friendly chocolate bars or candy coating. This type of chocolate holds a particular memory for many people in Poland, the country I live in. During the 1980s, it replaced authentic chocolate owing to shortages of essential ingredients.
Such a product typically contains no more than 7% cocoa [1]. In contrast, genuine chocolate has a cocoa content of at least 35% (typical bitter chocolate is 70%) [2]. Quite a difference, isn't it?
In my recent article, ‘From Numbers to Actions: Making Data Work for Companies‘, I delve into various issues, including the nature of insights. Many purported insights are similar to compound chocolate. At first glance, these so-called insights might seem authentic. But just like that piece of compound chocolate, when you take a closer look or ‘taste' them, they don't quite measure up.
Before I go any further, I shall stop and clarify, what I mean by the term ‘insight'.
What is an insight and why this term is so distinguished?
I explored various dictionary definitions of the word ‘insight.' The Oxford Learner's Dictionary defines it as ‘an understanding of what something is like‘. Dictionary.com describes it as ‘apprehending the true nature of a thing, especially through intuitive understanding.' Lastly, the Cambridge Dictionary defines it as ‘the ability to understand what something is truly like.'
The recurring theme here is ‘understanding' – understanding the essence of a matter. But what is this pivotal ‘matter'? It's something specific, profound, and capable of bringing notable shifts in the operations of our company.
Yet, our understanding of ‘insight' is still not complete. While it's about understanding a significant business matter, it also encompasses ‘actionability'. Genuine insights lead to concrete actions, driving decisions that embody the changes they suggest.
In essence: Insight is a profound understanding of a specific business matter that triggers decisions and actions.
Why are insights so important?
Today, the effective utilization of data is universally acknowledged as a foundation of competitive advantage. However, a surprisingly small fraction of companies harnesses its full potential. Merely 27% of companies view themselves as data-driven [3]. When envisioning businesses that truly leverage and derive value from data, the usual suspects come to mind: online giants like Amazon, Meta, and Netflix. Yet, the truth is, organizations of any size and across all sectors can fuel their growth through the efficient use of data.
Simply possessing data isn't enough. Even if the data is perfectly aligned with specific business needs, its true value emerges only when applied effectively. This means decisions and subsequent actions must be grounded in the trends, nuances, and insights gleaned from that data.
Numerous enablers can facilitate this process, but there are also significant roadblocks to consider. On the roadmap to success, several elements are vital: robustly designed and efficiently managed data sources, company leadership committed to data-driven decisions, a dedicated and well-trained data team, and the use of data-driven storytelling.
Why are there so few insights that align with this definition?
The overarching reason is the inherent challenge in crafting such insights. Luckily, there are tools in place that help companies navigate through this process. Consider the Insight Value Chain Model proposed by McKinsey [4]. This conceptual framework guides organizations in transforming raw data into actionable insights, thereby generating business value. The model shows four main steps that explain how data goes from being raw to making valuable decisions. In the picture below, I use an example of a retail company trying to boost its sales to show how this model works.
As illustrated above, the process is intricate. To gain a deeper understanding, let's dissect it systematically. Fundamentally, the pressing question we seek to address is:
How to increase the % of insight in the insight?
There are four fundamental characteristics that define an ‘authentic' insight. An insight that has 70% cocoa content, if not more, irrespective of whether it's bitter. I show them in the picture below:
An ‘authentic' insight:
- should give an understanding of a business matter…
- be specific…
- … and meaningful.
- should prompt decisions and actions.
Now, let's delve deeper into these characteristics.
How to produce an insight that gives understanding?
Authentic insights must be contextualized to maximize their impact and comprehension. Context enriches the narrative driven by data. Six methods to infuse insights with context include:
- Comparative context: compare product sales monthly or juxtapose actual costs against a budget or last year's same period.
- Scale Adjustment: highlight the cumulative impact over time or break down annual benefits into monthly or weekly gains for a more tangible perspective.
- Equivalence: aid comprehension by using familiar examples. Instead of stating: ‘your smartphone has 128GB storage', mention ‘it can store 25,000 photos*'.
- Historical context: display performance trends, considering seasonal or cyclical influences. Always compare whole periods.
- Informational context: offer details about patterns or anomalies without presuming correlation implies causation.
- Data validation: enhance trustworthiness by citing data sources, collection methods, and timeliness [5].
Secondly, never settle for the initial conclusion, especially when employing LLMs for analysis. Delve further until the conclusions resonate with genuine insight.
Third, trigger the spark. And the simpler tool you use, the more likely is that it will happen. Even if the technique employed seems simple. Remember what Archimedes said:
Give me a place to stand, and I will move the earth.
Below, I present some analyses executed using a basic tool like Excel. While these analyses are straightforward, they can yield valuable insights, potentially serving as a foundation for deeper exploration with more sophisticated programs or techniques.
The initial chart illustrates the fluctuations in the customer confidence index over a year, analyzed using Excel. As evident from the trend line and accompanying linear regression equation, the overarching trend is downward. Notable dips occurred during events like the C-19 lockdown and the outbreak of war in Ukraine. Currently, the trend is on an upward trajectory.
Another analysis, also conducted in Excel, aids in identifying peculiarities within the distribution of results. By employing a basic histogram chart, we can pinpoint outlying values and assess any irregularities in the frequency distribution. For instance, what might initially appear as a single distribution could, in reality, be three distinct ones, as demonstrated in the following example:
The final analysis, also performed in Excel, involves adding a trend line to a chart. This tool allows for the application of various functions, both linear and non-linear, along with the regression equation. Moreover, one can assess the fit's accuracy using the R-squared estimate.
How can we make insights more specific and meaningful?
Insights must be intrinsically linked to core business objectives and strategic initiatives. The stronger this connection, the less likely these insights will go overlooked.
Broadly, there are two types of indicators:
- KPIs (Key Performance Indicators)
- KCIs (Key Conceit Indicators).
If an indicator proves challenging to respond to, regardless of the magnitude of its change, it's likely a KCI – widely monitored within an organization but lacking in actionable value. Conversely, insights related to KPIs can instill a genuine sense of urgency, driving decision-making and action.
The closer a KPI aligns with corporate strategy, the more naturally it translates into tactical responses, as these are directly connected to pivotal business components.
KPIs must be deeply embedded within the company's DNA, spanning from top leadership to back-office employees. The balanced scorecard can be instrumental in disseminating targets and metrics across every division. By nurturing roles that seamlessly connect management, finance, and Data Science, a unified approach to target realization emerges. Emphasize business partnering across all areas of the organization, from sales to accounting. For suitable organizations, adopting agile management structures can enhance this integrated strategy.
How insight can prompt decisions and actions?
The initial step involves embracing the art of data storytelling. Communicating Insights should evolve beyond merely presenting intricate tables to decision-makers. Such an approach risks overwhelming them, prompting them to disengage.
Effective data storytelling stands on three tenets:
- Understanding context: recognizing what drives our audience.
- Employing narrative structure: implementing elements like the Storytelling arc [6].
- Utilizing effective visuals.
What constitutes an effective visual? Primarily, it should be clear and not confuse the audience. Hence, I advocate for the use of these three chart types:
For chart selection, use column or bar charts when comparing aggregated values like budget versus actual. A line chart is your go-to for analyzing trends. And if you're trying to understand how a part relates to the whole, a pie chart is ideal. These three chart types will likely cater to around 80% of your visualization needs unless there's a specific scenario like cohort analysis**.
When designing your charts, it's essential to eliminate any clutter. Remove extraneous elements like frames, support lines, and unnecessary data points, which can distract from the main message. Think of color and text as strategic tools in your arsenal; they should be used to emphasize and highlight key information, not just to beautify the chart.
Always be in tune with your audience. Test your visuals, see what works and what doesn't, and adjust accordingly. This iterative process is key to building a mutual understanding and ensuring your data tells a compelling story.
Lastly, ensure your narrative flows naturally. Avoid derailing your audience's attention with unnecessary and extensive suspense. Evaluate your storytelling using methods like the 3-minute story or the Big Idea [7]. For instance, I vocalize my narratives, be it articles or presentations. If I can articulate the story smoothly, it boosts my confidence in its resonance with the audience. Once you've won their attention, introduce key conclusions and call to action. Be sure to do that right after the story's climax – that's when they're most engaged and receptive. However, if reservations arise, prioritize active listening. Address any uncertainties, and if needed, suggest collaborative follow-up activities to foster understanding.
Conclusions
In this article, I walk through my method for crafting powerful insights. These aren't just any insights; they're the kind that guides businesses toward smart decisions. When used right, these insights can be game-changers, helping companies tackle tough situations or take advantage of great opportunities. Having the right data or the best tools isn't the whole story. How we share and explain these insights is just as crucial. It's all about making sure the message hits home, gets people thinking, and motivates them to act. In the end, the most valuable insights are those that lead to meaningful action and transformation.
*Assuming an average photo size of 5MB and 120GB of effective space on the smartphone
**Author's subjective assessment
[1] Wikipedia, Compound Chocolate
[2] Martinko, Katherine, What Does Cacao Percentage Mean on a Chocolate Bar?, February 6, 2021
[3] Szudejko, Michal, From Numbers to Actions: Making Data Work for Companies, August 14, 2023
[4] Hürtgen, Holger and Mohr, Niko, Achieving business impact with data, April 27, 2018
[5] Dykes, Brent, Contextualized Insights: Six Ways To Put Your Numbers In Context, October 18, 2018.
[6] Apple Podcasts, Narrative Arc: The Missing Tool in Your Data Stories with Brent Dykes, 2021.
[7] Nussbaumer Knafflic, Cole, Storytelling with Data, Wiley, 2015.
