How do you estimate price elasticity?
From a log-log model to an uncertainty interval: understand what a coefficient measures before simulating a price change.
12 min · Dataset MSC-001Try the tool ↗
Research & Evidence · Learn by doing.
Tools to analyze,
understand and make marketing decisions
A public laboratory for the methods, data and evidence that inform marketing decisions.
01 Explicit assumptions02 Downloadable data03 Visible limitations
Our evidence contract 01 / 04
From question to decision
A method has value only when it can be understood, reproduced and challenged.
The laboratory uses one consistent protocol for its reference dossiers. It shows what the data can establish — and what they cannot support.
View the featured dossiers ↓- 01Question
- 02Assumptions
- 03Data
- 04Model
- 05Calculation
- 06Uncertainty
- 07Limitations
- 08Decision
Scientific territories 02 / 04
The scientific and methodological hub
Six disciplines.
One standard.
Six fields for measuring effects, making models explicit and defining what the data can support in a decision.
01Δ
Marketing Measurement
Measure marketing effects while separating attribution, incrementality and causality.
- MMM
- Incrementality
- Experimentation
02ε
Pricing Science
Connect price, demand and perceived value with explicit assumptions and uncertainty.
- Elasticity
- Willingness to pay
- Conjoint
03C
Customer Science
Analyze customer value, retention and behavior over time without hiding data limitations.
- CLV
- Churn
- Cohorts
04μ
Market Research
Design reliable measures for observing markets, perceptions and preferences.
- Sampling
- Scales
- Bias
05ƒ
Marketing Models
Formalize response, diffusion and demand to compare mechanisms rather than narrate a curve.
- Saturation
- Carry-over
- Demand functions
06P
Decision Science
Compare scenarios and quantify uncertainty before preparing a human judgment.
- Monte Carlo
- Sensitivity
- Bayes
Featured dossiers 03 / 04
Start with a real question
Three methods.
Three bodies of evidence to build.
Dossiers designed as working objects: method, dataset, calculation and boundary of interpretation.
y = α(1 − e−βx)
When does marketing response saturate?
Compare functional forms, marginal returns and diagnostics before interpreting a curve.
10 min · Dataset MSC-002View the data ↘
τ̂ = (ȳT,post − ȳT,pre) − (ȳC,post − ȳC,pre)
Did a promotion create incremental sales?
Build a credible counterfactual and separate observed effect, true incrementality and timing shifts.
14 min · Dataset MSC-005View the data ↗
Method library 40 / 40
40 reference dossiers
From a question to a challengeable method.
Every dossier shows the statistical target, data, calculation, uncertainty, checks and decision boundary.
01MSC-P-001
Evidence foundations
How do you turn a marketing claim into a testable question?
Open dossier↗
02MSC-P-002
Evidence foundations
Correlation or causality: what can an analysis actually support?
Open dossier↗
03MSC-P-003
Evidence foundations
How should uncertainty in a marketing result be expressed?
Open dossier↗
04MSC-P-004
Evidence foundations
Statistical significance or effect size: which result should be interpreted?
Open dossier↗
05MSC-P-005
Measurement science
How do you measure a marketing construct that is not directly observable?
Open dossier↗
06MSC-P-006
Measurement science
How do you design and validate a measurement scale?
Open dossier↗
07MSC-P-007
Measurement science
Alpha or omega: how should scale reliability be assessed?
Open dossier↗
08MSC-P-009
Measurement science
PCA, EFA or CFA: which method should you choose?
Open dossier↗
09MSC-P-010
Experimentation and causality
When should you run a marketing experiment?
Open dossier↗
10MSC-P-011
Experimentation and causality
How do you design an A/B test that actually estimates an effect?
Open dossier↗
11MSC-P-012
Experimentation and causality
How many observations does an experiment need?
Open dossier↗
12MSC-P-013
Experimentation and causality
How do you measure campaign incrementality with a control group?
Open dossier↗
13MSC-P-017
Experimentation and causality
How do you detect selection, contamination and attrition in an experiment?
Open dossier↗
14MSC-P-018
Regression and econometrics
Predictive or causal regression: what are you trying to estimate?
Open dossier↗
15MSC-P-019
Regression and econometrics
How do you diagnose a marketing regression before interpreting it?
Open dossier↗
16MSC-P-022
Pricing science
How do you estimate price elasticity and its uncertainty?
Open dossier↗
17MSC-P-026
Choice models
Logit vs Probit: how do you choose for purchase probability?
Open dossier↗
18MSC-P-029
Customer science
Which customers have the highest probability of churn?
Open dossier↗
19MSC-P-027
Measurement science
TAM, UTAUT or UTAUT2: which framework should be used to study technology acceptance?
Open dossier↗
20MSC-H-001
Experimentation and causality
Measurement and causality: how can a marketing effect be established?
Open dossier↗
21MSC-H-002
Regression and econometrics
Marketing response models: shape, delay and saturation
Open dossier↗
22MSC-H-003
Pricing science
Pricing science: connecting price, demand and contribution
Open dossier↗
23MSC-H-004
Customer science
Customer and choice science: behavior, value and heterogeneity
Open dossier↗
24MSC-H-005
Measurement science
Measurement science: building valid indicators
Open dossier↗
25MSC-H-006
Evidence foundations
Statistical decision methods: choose, quantify, validate
Open dossier↗
26MSC-P-008
Measurement science
How do you validate a marketing measurement scale?
Open dossier↗
27MSC-P-014
Experimentation and causality
How do you design a marketing geo experiment?
Open dossier↗
28MSC-P-015
Experimentation and causality
How do you estimate an effect with difference-in-differences?
Open dossier↗
29MSC-P-020
Pricing science
How do you address price endogeneity?
Open dossier↗
30MSC-P-021
Regression and econometrics
Fixed or random effects: which panel model should you choose?
Open dossier↗
31MSC-P-023
Pricing science
How do you estimate a demand function?
Open dossier↗
32MSC-P-024
Pricing science
How do you simulate a price-volume-margin scenario?
Open dossier↗
33MSC-P-028
Customer science
How do you estimate CLV with BG/NBD and Gamma-Gamma?
Open dossier↗
34MSC-P-030
Customer science
How do you analyze retention with a survival model?
Open dossier↗
35MSC-P-031
Customer science
How do you build a useful customer segmentation?
Open dossier↗
36MSC-P-032
Customer science
How do you test segmentation stability?
Open dossier↗
37MSC-P-033
Regression and econometrics
How do you validate a marketing forecast?
Open dossier↗
38MSC-P-034
Evidence foundations
How do you build a Monte Carlo simulation for a marketing decision?
Open dossier↗
39MSC-P-035
Regression and econometrics
How do you model saturation and adstock?
Open dossier↗
40MSC-P-039
Evidence foundations
Which statistical test should you choose?
Open dossier↗
Open data laboratory 04 / 04
Reproducible datasets
Learn from data you can inspect.
Each example dataset is synthetic, documented and intentionally compact. It is designed to reproduce a method, not to generalize a result to a market.
FORMAT CSV UTF‑8 LICENSE CC0 1.0
Variables, methods and limitations →Price Elasticity
Price · volume · seasonality · 24 observations
Advertising Response
Spend · exposure · response · 24 observations
Promotion Incrementality
Test · control · sales · 24 observations
Customer Lifetime Value
Recency · frequency · value · 30 customers
Marketing Mix Modeling
Channels · sales · controls · 36 periods
Measurement Validity
Respondents · items · groups · 30 respondents
Monte Carlo Inputs
Scenarios · distributions · dependencies · 24 assumptions
Reproducible tools 05 / 05
MSC-T01 — MSC-T08
Five calculations, five interpretation boundaries.
Each tool exposes the calculation and states what it does not prove.
A/B sample size
3,841 per armTwo-sided normal approximation, 1:1 allocation, 80% power, α = 5%, without attrition or design effect.
Difference-in-differences
Difference-in-differences: +7.0 index pointsDescriptive 2×2 contrast; causal only when the design supports parallel trends.
Constant elasticity
-6.6%Conditional scenario, not a price recommendation.
Price-volume-margin
P₀ 100 € · VC 55 € · Q₀ 10 000
Contribution change: +3.8%Conditional scenario, not a price recommendation.
MSC-P-024 →Monte Carlo
Probability of positive contribution: ≈ 99.9% (MC SE ≈ 0.1 pp · Wilson 95% CI ≈ 99.6–100.0%)2,000 fixed-seed draws; only elasticity follows an unbounded normal, without MSC-010 dependencies.
MSC-P-034 →