beta // transparent output

MMM Scenario Laboratory

A transparent teaching laboratory for seeing how spend allocation and response-curve assumptions change an illustrative outcome range. It does not use your data and does not produce a forecast or causal estimate.

MODEL LAB / EXPERIMENT 001

Explore response curves without pretending they are your model.

This teaching laboratory uses fixed illustrative curves. Change the allocation, observe diminishing returns and uncertainty, and then inspect what a real MMM would need.

LOAD SCENARIO
WEEKLY MEDIA INPUTS$130,000
orders without modeled media contribution
ILLUSTRATIVE OUTPUTNOT A FORECAST
Modeled weekly outcome 1,103

938illustrative range1,268

Paid search

0
Paid social

0
Video & display

0
Theo, the Off Leash Marketer Australian Labradoodle

THEO'S MODEL NOTEBalanced spend still produces different response because each illustrative curve saturates differently.

WHAT THIS SHOWS

Carryover and saturation assumptions can change the apparent value of an additional dollar.

WHAT THIS DOES NOT SHOW

Your causal channel contribution, optimal budget, or a forecast grounded in observed business data.

WHAT A REAL MODEL NEEDS

A governed outcome, media history, controls, variation, time-aware validation, and uncertainty review.

What this laboratory teaches

The laboratory applies fixed illustrative response curves to three channel budgets. It demonstrates why additional spend can produce diminishing modeled response and why reallocating a fixed budget can change the central estimate and uncertainty range.

It is intentionally transparent: the curves are examples, not parameters learned from your company’s history.

Formula and worked example

Formula // inspect the definition

The illustrative response curve behind the laboratory.

Illustrative channel contribution Ceiling × (1 − e^(−spend ÷ half-saturation))

Spend produces a diminishing illustrative response as it approaches the fixed example ceiling.

Central modeled outcome Baseline outcome + sum of channel contributions

The channel examples are added to a user-entered baseline; none of these values are learned from observed company data.

Illustrative range Central outcome ± max(15% of outcome, 80% of incremental contribution)

A teaching boundary added around the example—not a statistical confidence or credible interval.

These fixed equations explain the interface. They do not estimate causality, forecast an outcome, or recommend a budget.

The three channel contributions are added to the baseline outcome. The illustrative range uses the larger of 15% of the modeled outcome or 80% of the incremental contribution on either side of the central value.

With the default weekly inputs—$60,000 paid search, $40,000 paid social, $30,000 video and display, and a 1,000-order baseline—the fixed example curves produce a central modeled outcome of about 1,103 orders. That is a demonstration of the math, not evidence about any real channel.

What it cannot tell you

The output is not a forecast, causal estimate, attribution result, or budget recommendation. It does not use observed outcomes, media history, controls, experiments, geography, seasonality, priors, or a fitted statistical model.

Use it to understand a modeling concept—not to approve a budget.

The next real step

Complete the MMM data-readiness worksheet and define the decision, outcome, row grain, historical window, media inputs, controls, and validation plan. Continue through the Build a Marketing Mix Model hub when the data contract is ready to review.

What this connects to

Turn the output into a rule your team can repeat.

Use the related hub, implementation notes, and templates to document the decisions behind the generated output.