Uplift#
The incremental effect of an action on an individual’s outcome.
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 Despite careful review, it may still contain mistakes, omissions, or out-of-date information.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#
Uplift measures the incremental impact of an action, treatment or intervention compared to not taking it. In machine learning, uplift modelling — also called incremental response or true lift modelling — predicts the change in outcome probability caused by applying a treatment (say, sending a marketing offer). It is about causal effect, not mere correlation.
The formula#
where \(Y\) is the outcome (purchase, churn, click), \(T\) is the treatment (1 = received, 0 = not), and \(X\) are the individual’s features.
A worked example#
For a promotional email, compare each customer’s treated and untreated purchase probability. Customer A: 30% with the email versus 25% without — an uplift of +5%, worth targeting. Customer B: 60% versus 60% — 0%, the email is irrelevant. Customer C: 20% versus 30% — a -10% uplift, meaning the email actively reduces the chance (a spam-sensitive recipient).
Why it matters#
Uplift buys targeting efficiency — spending only on customers who change behaviour because of the treatment — and cost savings, by skipping “sure things” who would act anyway and “sleeping dogs” who react badly. Used across marketing, healthcare, finance and policy, it is measured with the uplift (Qini) curve, its area AUUC, and the Qini coefficient.
Theme: Causal Inference & Uplift · All terminology
Hint
Mind map — connected ideas
Uplift Models · Uplift Score · Treatment Effect · Causal Inference · Conversion Rate Uplift · Qini Coefficient
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
See also
Source article Adapted (context, re-expressed) in our own words from: Uplift (insightful-data-lab.com).