How One Google Team Built Storytelling Into Analytics

Even the most advanced analytics models can fall flat if they don’t use the language of the organization’s decision makers. And when the related decision-making process stalls, it’s often because the data insights lack a clear narrative, business context, or connection to what executives care about. Those are lessons the author and her team learned at Google’s Small and Medium Business (SMB) division when her analytics team built a sophisticated model to optimize staffing for the company’s global support organization.

The model was able to forecast volatile demand across more than 100 countries by simulating thousands of possible scenarios and could recommend sales and customer support staffing levels. It accounted for seasonality, geographic differences, and even complex customer prioritization rules. The team validated the data, vetted assumptions, and pressure-tested the logic.

But when the team presented the model to senior stakeholders, they showed little enthusiasm. Instead of appreciating the model’s complexity, the stakeholders focused the discussion on the practicality of our recommendation model. One leader asked, “What does this mean for next quarter’s staffing in Latin America?” Another questioned how the recommendations would move the needle for her bottom line. Weeks of work stalled. No business decision was made.

This wasn’t an isolated incident. Repeatedly, technically sound models had failed to generate movement at the executive level. While many of the data professionals believed that better models would lead to improved business outcomes, executives were overwhelmed by the complexity and skeptical of black-box insights they could not contextualize.

The lesson was clear: Analytics must be built for how decisions are made, not just how data is analyzed. This required that we rethink the analytics stack — not just the data and models but also how storytelling can guide every stage of the process.

So in 2023, the team at Google began building an original framework with narrative at its core, with the goal of enabling business leaders to make faster and more confident decisions. The framework drew on lessons from a series of internal projects. Once the model was conceptualized and structured, it was refined and pressure-tested through mid-2024 with analytics team and business stakeholders. Since then, as head of analytics for Google’s SMB division, the author implemented this approach directly in high-impact projects involving sales strategy optimization, business planning, and executive decision enablement.

In this framework, each layer addresses a specific source of potential failure. Together, these four layers develop analytics that are based on how likely executives are to consume and act on data insights.

1. Data layer: Builds business-aligned data foundations.
2. Analytics layer: Designs flexible, interpretable models that anticipate and address executive needs.
3. Decision layer: Structures outputs around actionable levers, trade-offs, and constraints.
4. Narrative layer: Delivers coherent, contextualized insights that lead to action.
5. Each layer builds on the one below it to form a structured approach that helps executives make better, faster decisions.

show less