Analytical Writing: From Data to Insight

Data Do Not Speak for Themselves

A common misconception: "Data speak for themselves. We'll just show the numbers, and everything will become clear." Data do not speak — data are silent. It is the interpretation of data that speaks. And it is precisely in interpretation that the analyst creates or loses value.

Analytical writing is the transformation of raw data into an insight that demands action. Insight is not "sales grew by 12%." Insight is "sales grew by 12%, but margin fell by 8%, because growth was stimulated by discounts — this is not sustainable growth, and if we continue this policy, we will destroy profitability by Q4."

Structure of the Analytical Document

SCR (Situation — Complication — Resolution): a method developed at McKinsey. Situation — context shared by the audience (facts that both sides acknowledge). Complication — what has changed or presents a problem (tension). Resolution — recommended actions.

Example: "Our market has grown by an average of 8% per year over the past five years (Situation). However, in the last quarter, three key competitors entered with products priced 25% lower, which has already led to our share dropping from 34% to 29% (Complication). To protect our position, we need to review our pricing strategy within 90 days, accelerate the launch of the new product, and intensify work with loyal customers (Resolution)."

From Analysis to Recommendation

A frequent mistake by analysts: the text ends with analysis and no recommendation. "As the data show, the situation is ambiguous." A manager needs a recommendation. If you have conducted analysis and are not ready to recommend — analysis is not complete.

A recommendation should be: specific (what exactly to do), with time frames, with an assigned responsible person, and with a success criterion ("We will know the solution is correct when...").

Data Visualization

Data are often more convincing in visual form. Several rules: (1) one chart — one message (the headline should be an insight, not a description); (2) avoid 3D effects and decorative elements (they distort perception); (3) choose the chart type depending on the goal (line — trend, columns — category comparison, scatter plot — variable dependency).

Edward Tufte ("The Visual Display of Information"): maximum data with minimum ink — the principle of data-ink ratio. Remove everything that does not carry information.

Convincing Analytical Report

The best analytical reports from McKinsey, BCG, Morgan Stanley are not collections of data with comments. They are constructed as a narrative with an argument: there is a thesis, there is evidence, there is a conclusion. The reader never wonders "why do I need this?" — because every page answers this directly.

Question for consideration: Take the last analytical document you created. Does it contain an insight (not just data), a specific recommendation with a success criterion? Does it follow the SCR structure? What needs to be added or removed?