For subscriptions: every ad judged by who pays after the trial.
A trial costs little and proves less. Agents wait for trials to end, count who pays, and value each new subscriber at their first-year value, so an audience full of trials that never convert stops looking cheap.
The funnel, with each step’s worth
The subscription playbook’s starting funnel. You set the worths for your business. The suggested target is new subscribers under $60.
Where the numbers come from
Trials and subscriptions arrive through the events API, each subscription carrying its first-year value. Trial emails keep a holdout by default, so their lift is measured.
- 1Trial started$10
- 2Subscription startedits first-year value, about $200
What the agents do.
- 1
Pay for subscribers
Agents compare spend with the value of the subscriptions that follow, so an audience full of trials that never convert stops looking cheap.
- 2
Big moves wait for you
When an audience’s trials stop turning into subscribers, the agent lowers its bid within your limits and proposes moving its budget. Moving more than half of a campaign’s budget waits for your approval.
- 3
Convert trials by email
Lifecycle agents write to people in a trial within your contact limits and quiet hours, with a holdout so you can see what the emails changed.
- 4
Claims you can stand behind
The brand review agent checks every claim in an ad or email against your approved claims library. A claim without a source can’t be used.
- How they’re judged
- Subscriptions started and their first-year value, from your own conversion events.
- How often they act
- Daily if subscriptions arrive daily, weekly if they don’t.
- A good fit
- Subscription businesses that grow through paid acquisition.
- Not the best fit
- Subscriptions sold mainly through app stores, where the trial happens outside your own events.
Day 14: the trials end, and each ad is judged by who pays.
| When | The agents | You |
|---|---|---|
| Day 1 | One audience starts sending trials faster than any other: 41 in its first week, its most ever, at the lowest cost per trial of any audience. | Nothing needed. |
| Days 2 to 13 | The New Subscribers agent counts the trials but doesn’t move money on them. Its brief judges it on new subscribers, and none of these trials has ended yet. | Nothing needed. |
| Day 14 | The trials end. 2 of the audience’s 41 become paying subscribers, against 22 of 99 from everywhere else. Each counts at its first-year value, about $200, and a trial that never pays counts only as a trial. | Nothing needed. |
| Day 14, same morning | The agent lowers that audience’s bid, within your limits, and tells you why. | Read it in To read. |
| Day 15 | It proposes moving $30 a day from that audience to the two that convert best. | Approve or decline. |
Show it step by step
| When | The agents | You |
|---|---|---|
| Day 1 | One audience starts sending trials faster than any other: 41 in its first week, its most ever, at the lowest cost per trial of any audience. | Nothing needed. |
| Days 2 to 13 | The New Subscribers agent counts the trials but doesn’t move money on them. Its brief judges it on new subscribers, and none of these trials has ended yet. | Nothing needed. |
| Day 14 | The trials end. 2 of the audience’s 41 become paying subscribers, against 22 of 99 from everywhere else. Each counts at its first-year value, about $200, and a trial that never pays counts only as a trial. | Nothing needed. |
| Day 14, same morning | The agent lowers that audience’s bid, within your limits, and tells you why. | Read it in To read. |
| Day 15 | It proposes moving $30 a day from that audience to the two that convert best. | Approve or decline. |
Questions
What happens when trials spike but subscriptions don’t?
The agent judges on paid subscriptions, so an audience that only takes free trials doesn’t win budget. The report explains why, and moving more than half of a campaign’s budget waits for you.
What happens when you change your price?
Tell the agent in a sentence and update the subscription’s worth. Your guidance stays in its notebook and applies to every plan after.
What happens when someone is already subscribed?
The playbook’s audiences leave out current subscribers and anyone already in a trial.
Why value a subscriber by their first year?
It’s the playbook’s starting point, about $200 a year. Set the worth to whatever a subscriber is actually worth to you.
Can it judge on renewals?
Yes, if renewals arrive as a conversion event. Agents judge on the events in their funnel, each valued by what it’s worth.
Does it send the trial emails?
Yes, through your email platform, within your contact limits and quiet hours. Sending starts at You approve, and you can change that for each agent.
In the demo
Open the inbox in the demo (opens in a new tab)Spend where trials turn into subscribers.
Join the beta, or try the demo first.
