[ad_1]
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.
Source link
