🔗 Uplift Curve

Uplift Curve#

A curve showing cumulative incremental gain as more of the ranked population is treated.

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 Despite careful review, it may still contain mistakes, omissions, or out-of-date information.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#

An uplift curve is a tool for evaluating uplift models (incremental-response models). It plots the incremental gain — the extra benefit caused by the treatment — against the proportion of the population targeted, showing how much improvement you gain by targeting the top fraction of customers ranked by uplift score. It is closely related to, and often drawn as, the Qini curve.

Why not target everyone#

Traditional models (like logistic regression) predict response probability, but in marketing you don’t want to target everyone likely to buy — some would buy anyway, and some are even negatively influenced. The uplift curve reflects the model’s job of isolating persuadables, those who change behaviour because of the treatment.

How it’s built#

Rank customers by predicted uplift score from high to low, split them into buckets (top 10%, next 10%, …), and for each bucket compute the treated-minus-control difference in outcome rate. Plotting the cumulative incremental gains against the percentage targeted gives the curve: the X-axis is the fraction targeted, the Y-axis the incremental gain, with a random baseline line and the model curve ideally above it.

Reading it, with an example#

A steeper curve means the model is better at finding the most-influenced customers; a curve close to the random line adds little value; and the area between the model curve and the random line serves as a performance metric, much like AUC. For an email renewal campaign, the curve might show that targeting the top 20% by uplift score generates most of the incremental renewals, while targeting everyone simply wastes resources.

Theme: Causal Inference & Uplift · All terminology

Hint

Mind map — connected ideas

Qini Curve · AUUC (Area Under the Uplift Curve) · Uplift Score · Incremental Gain · Uplift Models · Uplift

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 Gain · Incremental Recovery Rate (IRR) · Incremental Revenue · Incremental Sales · Qini Coefficient