Subscription&ReplenishmentBrands
Cash flow vs ROAS conflict. Payback periods the platform cannot see. LTV projections that justify unprofitable acquisition. The economics are different here. The governance should be too.
This is what we see. This is how we think about it.
Sector Reality
What we observe in subscription accounts.
What breaks in this sector
- Cash flow pressure forces short-term ROAS targets that kill LTV
- Payback periods get ignored because the platform cannot see them
- First-order metrics look healthy while subscription churn accelerates
- Smart Bidding optimises for trial signups, not retained subscribers
Where Google Ads misleads
- ROAS calculated on first order, not customer lifetime
- Conversion tracking counts trial starts, not trial conversions
- Attribution ignores the months between acquisition and payback
- Campaign performance masks cohort-level retention decay
Why ROAS lies here
- A high first-order ROAS often correlates with low retention
- Promotional acquisition inflates ROAS while destroying unit economics
- The gap between acquisition and payback is invisible to Google
- LTV illusion: projected value used to justify unprofitable acquisition
Common agency failure
Reports on first-order ROAS while the business bleeds cash. Uses projected LTV to justify acquisition costs without tracking actual retention. Scales campaigns that look profitable in the platform while payback periods extend beyond cash flow tolerance. No visibility into cohort economics.
What big agencies optimise
- • Trial signup volume
- • First-order CPA and ROAS
- • Impression share on category terms
- • Month-over-month revenue growth
This weakens the business over time because cash flow erodes while metrics look healthy. The agency reports wins while the finance team sees losses. By the time the disconnect surfaces, months of acquisition spend have been misallocated.
Commercial Decisions That Matter
What we constrain, protect, and let scale.
What we constrain
- Acquisition spend that exceeds cash flow tolerance
- Campaigns optimising for trial volume without conversion visibility
- Promotional dependency that attracts low-retention cohorts
- PMAX consolidation before cohort signals are clear
What we protect
- Cash flow against payback period risk
- Cohort-level retention data as the source of truth
- Trial-to-subscription conversion rates by channel
- Margin contribution from retained subscribers
What we let scale
- Acquisition channels with proven payback under tolerance
- Campaigns targeting high-retention customer profiles
- Search terms with validated LTV correlation
- Remarketing to high-engagement trial users
Signals We Look For
What tells us risk is rising. What triggers intervention.
Risk indicators
- Payback period extending while first-order ROAS holds steady
- Trial-to-subscription conversion declining by cohort
- Cash deployed faster than receipts from previous cohorts
- PMAX claiming brand conversions while churn rises
- Promotional orders exceeding subscription orders
Intervention triggers
- Payback period exceeds cash flow tolerance
- Cohort retention drops below break-even threshold
- Trial conversion rate below 40% of historical average
- More than 25% of acquisitions from promotional campaigns
Fit Assessment
Who this is and is not for.
This applies if
- ✓Monthly ad spend above £40k with subscription or replenishment model
- ✓Commercial maturity to share cohort retention and payback data
- ✓Decision ownership sits with someone who understands cash flow risk
- ✓Willingness to constrain growth for sustainable unit economics
This does not apply if
- ✗Brands without at least 6 months of cohort retention data
- ✗Teams that report on first-order ROAS without payback visibility
- ✗Organisations where marketing and finance operate separately
- ✗Anyone seeking growth at any cost without cash flow scrutiny
Related Resources
Further reading for subscription brands.
Discuss whether this applies to your business.
A 30-minute conversation to understand your commercial context and whether our approach fits.
Book a Review→