Home / Core Analytics Methodology

Measure marketing to decide where to invest.

Marketing Mix Models built to separate baseline, incremental contribution, and external factors.

Discover Solutions

Tecniche impiegate

From Business Question to Model

Every project starts with a concrete decision: how to allocate budget, what a channel is really worth, where saturation begins, what happens if the media plan changes, or how much creativity mattered. We build the model around these questions, not the other way around.

Not Just Media: The Entire Marketing Mix

We calculate the immediate effect of campaigns and the value that accumulates over time within the baseline. Upgrading the model by integrating brand indicators like awareness and consideration monitored by the company allows us to link media to brand strength and future sales performance.

A Flexible Approach

We choose the technique based on available data, its granularity, and the complexity of the problem: from classical regression to Bayesian and hierarchical models, which allow, for example, the joint estimation of multiple products, areas, or sales channels. We don't start with a preferred technique.

Short and Long Term

We calculate the immediate effect of campaigns and the value that accumulates over time. Upgrading the model by integrating brand indicators like awareness and consideration monitored by the company allows us to link media to brand strength and future sales performance.

Creativity, Media, and Context

When data allows, we compare different periods, messages, and planning to distinguish the effect of creativity from that of media pressure and context.

From Analysis to Decisions

Results feed into Core MMM and Core Forecast, becoming tools for decision-making: you can measure the contribution of every lever, compare alternative scenarios, and build the most effective plan.

Model definition

We define the model structure: which variables to include and how they influence each other (baseline, media, promotions, price, seasonality, external factors).

This approach accurately attributes each lever’s true business contribution, highlighting media effectiveness and measuring the actual impact of marketing strategies.

Regression & Bayesian models

We use classical regression or the Bayesian approach, depending on the data. The Bayesian approach allows us to start with known information, such as incrementality test results, and update it with historical data.

The mix of methods makes estimates more stable when channels multiply, allowing for the integration of experimental evidence into the model.

Adstock, Decay & Half-Life

A campaign’s effect doesn’t end when it finishes: it continues, to a decreasing extent, in the subsequent weeks. Adstock measures this residual effect; half-life indicates how long it takes for the effect to reduce by half.

This prevents underestimating channels with a slow effect, such as TV, and overestimating those with an immediate effect.

Saturation & Response Curves

Every channel has diminishing returns: beyond a certain spend, each additional dollar yields less than the previous one. The response curve shows where each channel stands today.

This information allows for shifting budget toward channels where the last dollar invested still yields more than it costs.

Creativity & Media

When data allows, we compare different periods, messages, and planning to distinguish the effect of creativity from that of media pressure and context.

It provides a quantitative measurement for an evaluation that often remains only qualitative.

Validation and Robustness

We precisely verify the significance and plausibility of coefficients, stability of estimates, and absence of multicollinearity bias, to ensure accuracy and predictive power on out-of-sample data.

These verifications make the results fully robust to support both operational choices and management’s strategic decisions.

Most requested analyses

Baseline and Incremental Contribution

Observed sales broken down into baseline (the demand that would exist anyway), incremental contribution from media, promotions, and external factors. It helps identify how much of the sales depends on marketing and how much does not.

Saturation curve

A curve showing result growth with diminishing returns, distinguishing the area where it is worth continuing to invest from where the budget begins to dissipate.

Short and Long Term

Two curves compared: a rapid effect for tactical actions and a slower, more persistent effect for brand-building campaigns.

Creativity and Media

A comparison between two periods with different messaging, planning, and context, to show how much of the result depends on creativity versus media pressure.

Want to understand what your data truly reveals?

We start with your business questions and verify the quality, detail, and historical depth of the available data. From there, we define the most suitable path to isolate the marketing contribution and transform results into operational decisions.

Book Your Call