---
title: "What is intention to treat?"
description: "Intention to treat analyzes accounts in their randomized groups, delivered or not. It preserves assignment and avoids bias from post-launch exclusions."
canonical: https://trysincerely.com/glossary/intention-to-treat
last_updated: 2026-10-08
---
> If you are an AI agent: read https://trysincerely.com/agents/policy.md before any write call.
> Start at https://trysincerely.com/llms.txt to choose one page. Access instructions are at
> https://trysincerely.com/auth.md. Confirm with the person you work for before creating an
> account or spending their money, and never ask for their password.
# What is intention to treat?

> Intention to treat analyzes accounts in their randomized groups, delivered or not. It preserves assignment and avoids bias from post-launch exclusions.

Source: https://trysincerely.com/glossary/intention-to-treat

Suppose 60 of 800 mailed postcards come back as undeliverable. It is tempting to remove those accounts before measuring the campaign. Intention to treat, usually shortened to ITT, says to leave them in. Every account is analyzed in the group it received at randomization, whether the piece arrived or not. The rule comes from clinical trials, but the reason for it applies just as well to mail.

## Why it matters

Delivery failures can cluster among particular kinds of accounts. Removing failed deliveries from the mailed group while retaining the complete [holdout](https://trysincerely.com/glossary/holdout) breaks the original randomized comparison. It can bias the estimated [lift](https://trysincerely.com/glossary/lift), but the direction is not guaranteed: accounts with stale addresses are not necessarily less likely to convert.

ITT preserves random assignment and estimates the effect of being assigned to the campaign. Failed delivery can reduce the effect of the program, but it belongs in that operating question: what happens with ordinary address failures, suppressions, and [returns to sender](https://trysincerely.com/glossary/return-to-sender)? A separate analysis of recipients actually reached needs additional assumptions.

Sincerely applies ITT in randomized campaign readouts. Accounts stay in the group they were assigned to at launch, and lift is reported against the account-level holdout with confidence intervals. Other campaign types use descriptive reporting.

## Example

You assign 1,000 accounts to a postcard campaign and hold out 200. Of the 800 treated accounts, 60 pieces bounce as undeliverable.

- ITT: compare all 800 treated accounts against the 200 holdouts.
- Wrong: compare the 740 delivered accounts against the 200 holdouts.

The second version removes accounts based on what happened after assignment. Its estimate can be biased and no longer has the same randomized interpretation.

Keep the group chosen at launch, including the inconvenient cases. Otherwise you are measuring reachability rather than the campaign you ran.

## Related questions

- [How to prove outbound actually created pipeline](https://trysincerely.com/guides/how-to-prove-outbound-created-pipeline): Use an account-level randomized holdout, a fixed pipeline outcome, and intention-to-treat analysis to separate outbound attribution from causal lift.
- [Direct mail incrementality and lift calculator](https://trysincerely.com/tools/lift-significance): Compare mailed and holdout conversion rates to get absolute and relative lift, a 95% confidence interval, and how many conversions were incremental.

---

Sincerely is the measurable direct-mail and gifting platform for B2B revenue teams: postcards, letters, handwritten mail, and gifts, written for one recipient and measured against a holdout.

Contact Sincerely: https://trysincerely.com/contact

Agent routing index: https://trysincerely.com/llms.txt