Most models predict who will convert. That is the wrong question for a
marketing budget. What you actually want is who converts only because you
reached them, the incremental lift, not the raw likelihood. This is an
uplift model: it estimates the causal effect of sending an email on
each customer, using T-learner and X-learner meta-learners validated
with the Qini curve. It runs on a real randomized experiment, the
Hillstrom email dataset
of 64,000 customers.
⚠️ Demo is running on placeholder coefficients. Run
causal-uplift/train.py to populate it with values learned from real data.
Map a customer
Adjust the customer and watch them move across the persuadability map.
Left to right is how likely they are to visit anyway; bottom to top is the causal
lift from emailing. The corner they land in is the decision.
Persuadablesemail tips them over
Sure Thingsconvert anyway
Lost Causeswon't convert either way
Sleeping Dogsemailing backfires
baseline visit likelihood, without email →
causal uplift from emailing ↑
-
-
-
without email
-
with email
-
uplift (causal)
Does the model actually rank uplift?
-
Average lift email vs no email
-
Qini T-learner
-
Qini X-learner challenger
-
Customers randomized trial
Qini curve: incremental visits captured vs. fraction targeted. Above the diagonal means the model beats targeting at random.Actual visit lift by predicted-uplift decile. Monotonic and rising means the ranking is real, not noise.Predicted uplift across all customers. Some are genuinely persuadable, some barely move, a few are negative.The raw experiment: emailing lifts the average visit rate. Uplift modeling turns that average into a per-customer decision.
How it works
A real experiment. Customers were randomly assigned to receive an email or not, so a simple difference in visit rates is already an unbiased causal effect. Randomization is what makes the uplift identifiable.
T-learner. Two models are fit: one on the treated group, one on the control group. The predicted uplift for a customer is the difference between what each model says, P(visit | emailed) minus P(visit | not emailed).
X-learner. A stronger challenger that imputes each customer's individual treatment effect and models it directly, which helps when the treated and control groups are unbalanced. Reported alongside as a Qini benchmark.
Qini curve. Accuracy is meaningless here (you never see both outcomes for one person). Instead the model is judged on whether ranking customers by predicted uplift captures more incremental visits per email than random, which is exactly what a Qini curve measures.
Decision. Each customer lands in one of four quadrants. The budget goes to Persuadables, and Sleeping Dogs get left alone.