How Vent Now A/B tested its app against iMessage and cut churn by 40%
"Nobody wants to install more apps. Everyone has iMessage, and texting is more familiar." — Jack Retterer, co-founder and CEO, Vent Now
| 20% | more messages per session on iMessage than in the app |
| ~40% | more daily activity per user |
| 40% | lower churn on weekly plans |
| 16% | higher lifetime value |
Company: Vent Now
Product: Aspen, an AI companion for emotional growth
Category: Consumer subscription
Summary
Vent Now's product is a conversational AI friend and therapist named Aspen. Previously, their primary UI was in a branded app.
Vent Now ran an experiment putting part of its user base on iMessage keeping everything else, back-end, model, and harness, the same with exceptional results.
Users on iMessage sent 20% more messages per session, were about 40% more active on a daily basis, and churned 40% less on paid plans. Early results estimate a 16% higher lifetime value.
Background
Vent Now is building Aspen, an AI companion somewhere between a friend and a therapist. Users talk to her about job interviews, stress, and the daily frustrations in life.
In this post we will walk through why Vent Now ran the test, the experiments results, and what the results do not prove.
Why they ran the test
"The more human we make Aspen seem, the more effective she is at being a therapist." — Jack Retterer
VentNow started testing iMessage as a channel to make Aspen feel more human.
Their working theory is that Aspen's usefulness scales with how human she feels, and a natural message thread inside iMessage where humans already interact with each other does that best.
Vent Now's plans carry usage limits with paywalls. Making the product feel more human made users talk more, getting more value out of the product & hitting limits faster.
Results of the Experiment
Vent Now measures usage in sessions. A user opens a conversation and stays in it, sometimes for an hour.
Messages per session: 20% higher on iMessage.
Daily activity: about 40% higher. Vent Now expects this to continue going up once proactive messaging ships.
For example, if a user mentions a job interview on Tuesday, Aspen can send a text on Thursday morning, something a friend would do.
Churn on weekly plans: 40% lower.
Lifetime value: 16% higher. Their plans carry usage limits. Users who send more messages hit those limits sooner, and a meaningful part upgrade or buy more credit. More interactions means more upgrade events.
What these numbers do not show
Four caveats, all of which Vent Now stated plainly.
The cohorts were not randomized. iMessage access rolled out by price tier, starting with the highest and working down. Higher-tier subscribers retain better and spend more before any interface is involved; however, these numbers are head to head against their user type. IE live time value is 16% higher than the same tier customers not using iMessage.
The conversion effect is upgrade conversion, not signup conversion. No one signs up for Vent Now through iMessage today. Every user in this comparison onboarded in the app or on the web and then enabled texting, so the test says nothing yet about whether iMessage acquires users more efficiently. Vent Now is running two onboarding experiments now, one routing web signups into a free iMessage trial and one starting from an inbound message. Those will produce their first top-of-funnel comparison.
One product, one category. This is a single test at a single company in a category.
Results
| Metric | iMessage vs. in-app |
|---|---|
| Messages per session | +20% |
| Daily activity per user | ~+40% |
| Churn, weekly plans | -40% |
| Lifetime value | +16% |
Vent Now found Linq through Tomo. They noticed Tomo's product, worked out what was powering the thread, and traced it back to us.
Next on their roadmap: proactive messaging, in-thread payments so a user can upgrade without leaving the conversation, and onboarding that begins with a text rather than a download. All powered by Linq.
What we would take from this
Native messaging formats notably increase the usage of conversational AI While this is only one companies results, the direction and magnitude of the impact are significant. Most product teams run an A/B testing engine to improve conversion a few percent. Switching to a native conversational user interface has a far greater impact.
Test interface against interface. Vent Now held the backend constant and changed the surface, which is why the comparison means something.
An interface people already use removes work you would otherwise do in product. Nobody had to be taught how to text.
Linq is the conversational messaging API for iMessage, RCS, SMS, and voice. 99.95% uptime backed by an SLA, sub-120ms latency, SOC 2 Type II certified, and no per-message pricing.
