Beyond Qualified Trials

How Rise achieved 30% higher ROAS on Meta with predictive value signals

Rise already combined high-velocity creative testing with sophisticated optimisation and internal predictive modelling. Day30 built on this foundation with predictive value signals designed for advertising optimisation. In a controlled Meta experiment, the approach delivered 30% higher net D33 ROAS than Rise’s existing Qualified Trial baseline.

+30%

Net D33 ROAS vs Qualified Trial

+23%

Trial-start rate vs Qualified Trial

Controlled A/B Test

Using Meta experiments

All results are relative or indexed, with no absolute spend or revenue disclosed. Net D33 cohort ROAS was measured using consistent MMP last-touch attribution and backend net revenue, after fees and refunds.

Rise at a glance

Rise at a glance

Company

Company

Rise

PRODUCT

PRODUCT

A consumer subscription app helping people improve their sleep

A consumer subscription app helping people improve their sleep

CHANNEL

CHANNEL

Meta

Meta

OBJECTIVE

OBJECTIVE

Improve revenue efficiency beyond an established Qualified Trial signal

Improve revenue efficiency beyond an established Qualified Trial signal

MEASUREMENT

MEASUREMENT

Net D33 cohort ROAS using attributed backend revenue

Net D33 cohort ROAS using attributed backend revenue

The acquisition challenge

Rise’s Qualified Trial event already offered a more selective indication of customer quality than a conventional trial start.

The opportunity was to go further: could a signal based on predicted revenue help advertising campaigns distinguish not just likely subscribers, but users expected to generate more value?

THE DAY 30 APPROACH

While Qualified Trial identified promising trial behaviour, Day30 used machine learning to estimate differences in expected net customer value.

Those predictions informed an early signal specifically designed for advertising optimisation.

To evaluate the approach fairly, the teams used the same acquisition cohorts and consistent MMP last-touch attribution linked to backend net revenue, rather than relying solely on in-platform ROAS. The predictions also considered refunds when estimating net customer value.

THE PILOT

Rise tested Day30’s approach against its established Qualified Trial optimisation in a controlled Meta experiment.

Creative, budgets and targeting were held consistent, with the optimisation approach as the deliberate variable.

The experiment ran for over a month, with additional time for cohorts to mature to D33 before the final comparison.

The result

Day30 delivered 30% higher net D33 ROAS than Qualified Trial.

The public comparison below is indexed to Rise’s existing baseline:

Optimisation signal

Net D33 ROAS (indexed)

Day30 predictive value signal

130

Rise Qualified Trial

100

Day30’s campaign also achieved a 23% higher trial-start rate and a lower observed refund rate. Although initial acquisition costs were higher, stronger downstream customer value more than offset them, resulting in 30% higher net D33 ROAS.

“We could see the difference in the quality of trials we were acquiring, even with the same creative. Higher conversion rates meant we were reaching more valuable users, giving us confidence to expand the approach into more campaigns.”

Jeff Kahn, Rise

The three pillars

What made the approach different

01

Designed for advertising optimisation

Rather than treating predicted value only as a reporting metric, Day30 used it to distinguish users by expected net revenue.

02

Revenue-aware prediction

The approach considered net customer value, including refunds, when developing its predictive signals.

03

Measurement tied to outcomes

Matured cohorts and consistent MMP attribution linked to backend net revenue enabled a like-for-like comparison.

  • 01

    Designed for advertising optimisation

    Rather than treating predicted value only as a reporting metric, Day30 used it to distinguish users by expected net revenue.

  • 02

    Revenue-aware prediction

    The approach considered net customer value, including refunds, when developing its predictive signals.

  • 03

    Measurement tied to outcomes

    Matured cohorts and consistent MMP attribution linked to backend net revenue enabled a like-for-like comparison.

  • 04

    Trusted evaluation

    Opal's own net-revenue reporting and a controlled test verified the result in Opal’s own net-revenue reporting.

  • 01

    Designed for advertising optimisation

    Rather than treating predicted value only as a reporting metric, Day30 used it to distinguish users by expected net revenue.

  • 02

    Revenue-aware prediction

    The approach considered net customer value, including refunds, when developing its predictive signals.

  • 03

    Measurement tied to outcomes

    Matured cohorts and consistent MMP attribution linked to backend net revenue enabled a like-for-like comparison.

  • 04

    Trusted evaluation

    Opal's own net-revenue reporting and a controlled test verified the result in Opal’s own net-revenue reporting.

  • 01

    Designed for advertising optimisation

    Rather than treating predicted value only as a reporting metric, Day30 used it to distinguish users by expected net revenue.

  • 02

    Revenue-aware prediction

    The approach considered net customer value, including refunds, when developing its predictive signals.

  • 03

    Measurement tied to outcomes

    Matured cohorts and consistent MMP attribution linked to backend net revenue enabled a like-for-like comparison.

  • 04

    Trusted evaluation

    Opal's own net-revenue reporting and a controlled test verified the result in Opal’s own net-revenue reporting.

FROM PILOT TO WIDER ROLLOUT

Following the pilot, Rise continued working with Day30 and began expanding the predictive-value approach into more app acquisition campaigns, including scaling and creative testing.

The rollout is ongoing. For subscription app growth teams, the takeaway is clear: even with strong creative and sophisticated optimisation already in place, refining the signal guiding ad delivery can unlock meaningful additional performance.

FROM PILOT TO WIDER ROLLOUT

Following the pilot, Rise continued working with Day30 and began expanding the predictive-value approach into more app acquisition campaigns, including scaling and creative testing.

The rollout is ongoing. For subscription app growth teams, the takeaway is clear: even with strong creative and sophisticated optimisation already in place, refining the signal guiding ad delivery can unlock meaningful additional performance.

Ready to see if predictive signal engineering can improve your ROAS on Meta?

© 2026 Day30. All Rights Reserved.
© 2026 Day30. All Rights Reserved.
© 2026 Day30. All Rights Reserved.