Test an outbound message and the audience rule behind it

The flow under test

The rule has two conditions, and the test checks each one failing as well as both passing.

Testing both negative combinations is what proves the rule rather than the message. A message shown on every page would satisfy a positive-only check completely, and nobody files a bug about seeing a message they were not meant to see.

The test file

Each combination is asserted by the count of messages shown rather than by the presence of text, which makes a hidden message and an absent one the same verifiable thing.

---
mode: testing
url: https://my-testing-repo-main.vercel.app/intercom-clone-app/outbound?reset=true
max_steps: 40
tags: [intercom, support, marketing]
---

# Intercomm 31.4: Outbound message display

Catalog objective: verify a targeted outbound message shows on the matching page.
Key assertion: the message displays with a working call to action.

## Verify the message shows on the matching page
With "Pricing page" selected and the plan on "Pro", verify "Messages shown on this page" reads 1 and the visitor view shows "Not sure which plan fits?" with the body "Book 15 minutes with us and we will size it with you."

## Verify it does not show on a non-matching page
Select "Home page" in "Page they are on" and verify "Messages shown on this page" reads 0 with the note that no outbound message targets this page and plan.

## Verify the audience rule is respected
Select "Pricing page" again, then select "Free" in "Their plan", and verify "Messages shown on this page" reads 0.

## Verify the call to action works
Select "Pro" in "Their plan", click the "Book a call" button, and verify a green banner reads ""Book a call" clicked - the click is recorded against Not sure which plan fits?" and the CTA clicks card shows "Clicks" of 1.

The final step clicks the call to action and asserts the click is attributed to the right message, since an unattributed click is a conversion nobody can report on.

Step by step coverage

A match, two misses, and a conversion.

Changing one condition at a time across steps two and three is what makes each half of the rule independently verifiable.

Running it

kane-cli testmd run tests/intercom/outbound-message_test.md

Add a repeat visit, since a message that reappears after being dismissed is the fastest way to make a helpful prompt feel like spam.

Targeting rules are configured by marketers and rarely regression tested. Running Kane CLI in CI covers catching a rule change, and the Kane CLI introduction covers getting started.