🔗 Incremental Gain

Incremental Gain#

The improvement in outcome from targeting versus a baseline.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

Incremental gain is the extra positive outcome — conversions, sales, sign-ups — produced by applying a treatment (a campaign, discount or intervention) versus not applying it, measured for a specific slice of the population. It is the building block of the cumulative incremental gain and the Qini curve: the step-by-step value added as you target more people.

The formula#

For one segment,

\[\text{Incremental Gain} = \text{Responses}_{\text{treatment}} - \text{Responses}_{\text{control}},\]

or, in probability terms, \(\left(P(\text{Outcome} \mid \text{Treatment}) - P(\text{Outcome} \mid \text{Control})\right) \times n\), where \(n\) is the number of individuals targeted in that segment and the control response is what the same group would have done untreated.

A worked example#

Rank 2,000 customers by uplift score and split into deciles of 200. In the top decile, the treatment group converts at 18% (36 purchases) against a control of 10% (20) — an incremental gain of 16. The second decile converts at 14% (28) versus 12% (24) — a gain of 4. Their running total, the cumulative incremental gain after the top 20%, is 20 extra purchases.

Where it fits#

Incremental gain is the local effect in each segment; summing it as you move down the ranking gives the cumulative incremental gain, and plotting that against the fraction targeted traces the Qini curve. Reading it segment by segment shows where uplift is strongest and where to stop targeting — once the gain flattens or turns negative, additional targeting destroys value.

Theme: Causal Inference & Uplift · All terminology

Hint

Mind map — connected ideas

Cumulative Incremental Gain (CIG) · Total Incremental Benefit (TIB) · Qini Curve · Conversion Rate Uplift · Treatment Effect · Incremental Revenue

Hint

More in Causal Inference & Uplift

AUUC (Area Under the Uplift Curve) · Causal Effect · Causal Impact · Causal Inference · Causal ML (Causal Machine Learning) · Causal Trees · Cumulative Incremental Gain (CIG) · Cumulative Uplift · Incremental Conversions · Incremental Recovery Rate (IRR) · Incremental Revenue · Incremental Sales · Qini Coefficient · Qini Curve