TL;DR:
In this first installment of the Measuring Marketing series, we explore different marketing measurement approaches and how to evaluate the best method for your business needs. Top-down methods use aggregated data to capture broad trends and evaluate strategic actions such as optimizing budget allocations. Bottom-up methods rely on granular user-level data to measure specific actions like ad impressions or conversions. In order to help you decide the best measurement approach, we recommend evaluating your capabilities in the PAD framework—Problem, Ability, and Data. Ultimately, effective measurement depends on understanding your goals, capabilities, and available data.
Not all insights are created equal — there is no one-size-fits-all approach in how you measure your business success. Choosing the right measurement approach is key to turning data into decisions that drive results and reduce risk. In this post, we’ll explore two core types of measurement: top-down and bottom-up. We further discuss how to evaluate both your data and your broader business context to align on your analytics journey.
Top-Down Approach: A View from Above
The top-down approach starts with aggregated data — such as time series (e.g. daily sales volume over time or before/after an event) or measures across geography (geospatial data; e.g. sales across different markets). It’s commonly used to understand how marketing, pricing, or strategic decisions play out over time or across regions, while accounting for external influences like seasonality or market-wide shifts. This approach is best suited for decision-makers aiming to guide broad strategies rather than changing individual behavior.
For time series data, methods like media mix models, synthetic control, and impulse response analysis are commonly used. While they vary in complexity, these approaches help practitioners calculate key metrics such as incremental or marginal return on investment (RoAS). These insights guide scenario planning, budget optimization, and validating strategic decisions. In geospatial analysis, tools like opportunity analysis, drivetime segmentation, and gravity models help prioritize which markets to invest in based on strategic potential.
Bottom-Up Approach: Trends from Below
The bottom-up approach focuses on the most granular level of data — individual people and actions. This includes user impressions, clicks, website visits, and conversions like purchases or donations. The approach also requires a high degree of control over data tracking and execution — making it most effective when organizations can directly influence individual-level experiences in near real-time.
Data infrastructure commonly used for this approach includes multi-touch attribution, customer journey analysis, and user-level experimentations. Common measurement strategies involve A/B testing, difference-in-differences, and regression discontinuity designs. These analyses evaluate the impact of targeted strategies—such as campaigns or segments—to optimize tactics like retargeting, frequency capping, and personalized messaging.
How Do You Decide? PAD Framework
Choosing the right measurement approach depends on more than just preference — it requires alignment with your business context. The PAD framework — Problem, Ability, and Data — helps guide that decision by clarifying what you’re trying to solve, what actions you can take, and what data you have.
Start by defining the problem: What are decision-makers trying to learn? What gaps exist in your current measurement or strategy? Are you solving for long-term strategy, short-term tactics, or both? This step helps set clear expectations around the purpose and scope of the analysis.
Next, assess your ability to act on the insights. Can your team implement changes at a user level (e.g. targeted ad delivery or apply frequency caps on re-targeted ads), or are decisions made at a broader level (e.g. budget allocations or market planning)? If you take more granular and tactical actions, the bottom-up approaches are recommended.
Finally, consider your data: Is it granular, timely, and clean enough to support user-level measurement, or is it aggregated and better suited to top-down analysis? Prioritize approaches that align with what your current systems can realistically support, while also identifying gaps to address for future improvement.
Beyond the Measurements: Using Judgement and Context
All PAD exercises will leave you to one universal takeaway: No measurement approach is going to be perfect. Every method comes with trade-offs, assumptions, and blind spots. Choosing the right approach means balancing precision with practicality, and using insights as tools to inform decisions — not as truth-teller.
All PAD exercises will leave you to one universal takeaway: No measurement approach is going to be perfect.
In the next article, we’ll dive into measuring incrementality—a focused application to test what truly drives impact. This controlled and guided approach supplements both top-down and bottom-up approaches to remedy their methodological weaknesses.
Guiding Your Analytics Journey with the Right Measurement
We can help guide you through your analytics journey — from clarifying the problem to identifying the right approach using the PAD framework. Whether you’re refining strategy, evaluating performance, or navigating data complexity, we’re here to turn insights into impact.
