Subscription DTC: Why First-Order ROAS Negative Is Actually Healthy
For replenishment and subscription brands, a 0.8x first-order ROAS often hides a healthy 3x LTV-adjusted return by month 4. Here's when it's the right strategy — and how to prove it.
Quick answer
If your AOV is roughly your CAC and customers reorder on a predictable cycle (30-90 days) at 55%+ M2 retention, a first-order ROAS of 0.7x–0.9x is healthy — not broken. The channel pays back on refill 2 or 3, and your LTV-adjusted ROAS at month 4 typically lands between 2.5x and 3.5x. Kill the channel only if M2 retention drops below 45% or refill cadence slips by more than 20%.
Negative first-order ROAS (subscription DTC)
A paid-acquisition state where first-purchase revenue is below ad spend, but cohort LTV recovers spend within 2-4 refill cycles.
In subscription and replenishment commerce, first-order ROAS measures only the revenue from a customer's initial purchase against the ad spend that acquired them. For brands where the real margin lives in refills — coffee beans, skincare serums, supplements, pet food — a ROAS below 1.0x on order one is often the deliberate price of acquiring a recurring customer. The relevant question is not whether order one is profitable, but how quickly the cohort's cumulative contribution margin overtakes the original CAC. That curve, not the first-order number, is what determines whether the channel is healthy.
Most CFOs look at a Meta dashboard, see 0.8x ROAS, and ask why the channel is still on. It's a reasonable question if you sell one-shot products. It's the wrong question if you sell something a customer buys again in 30 days.
The mechanism is simple: paid acquisition pays for the first transaction; refills pay for everything after. If refill behavior is predictable enough to forecast, the first order is just the entry fee.
Why first-order ROAS lies for subscription brands
A one-time purchase business and a subscription business can have identical first-order ROAS and completely different P&Ls. The one-time business has booked its entire customer relationship in that one transaction. The subscription business has booked roughly 15-25% of it.
This is why the question is closer to CAC payback period than to ROAS. You're asking: how many refill cycles until cumulative gross margin per customer crosses CAC? For most well-run replenishment brands that number is 2-4 refills, which on a 30-day cadence means month 3 to month 5.
The two numbers that decide it
Two inputs separate healthy negative ROAS from a slow bleed: month-2 retention (the % of acquired customers who take refill 1) and refill cadence drift (how much your actual refill interval slips vs. the planned cycle). If M2 retention is above 55% AND cadence slip is under 15%, sub-1x first-order ROAS is almost always defensible. If either breaks, the math inverts fast.
Modeling the payback curve
Build the curve at the cohort level, not the campaign level. For each weekly or monthly acquisition cohort, track cumulative contribution margin per acquired customer against the original blended CAC for that cohort. The intersection point is your real payback period.
Most teams overcomplicate this. You need four columns: cohort week, customers acquired, CAC per customer, and cumulative contribution margin per customer by month. Plot the latter as a curve; the month it crosses CAC is the payback month.
The CFO conversation gets easier when you can show last quarter's cohort already crossed the line on the timeline you predicted. That's the artifact — a chart with actuals catching up to the model — that defends the channel. There's a deeper walkthrough in our companion piece on modeling the month-4 subscription payback curve for the CFO.
Healthy ranges by subscription vertical
Typical first-order ROAS, M2 retention, and LTV-adjusted ROAS at month 4 by vertical
| Vertical | Refill cycle | Healthy first-order ROAS | M2 retention | LTV-adjusted ROAS (M4) |
|---|---|---|---|---|
| Coffee subscription | 21-30 days | 0.7x – 0.9x | 60-70% | 2.8x – 3.6x |
| Skincare replenishment | 45-60 days | 0.6x – 0.9x | 55-65% | 2.5x – 3.2x |
| Supplements / vitamins | 30 days | 0.8x – 1.0x | 55-65% | 3.0x – 4.0x |
| Pet food | 60-90 days | 0.9x – 1.1x | 65-75% | 2.4x – 3.0x |
| Snack / treat box | 30 days | 0.8x – 1.0x | 45-55% | 2.0x – 2.6x |
Refill cadence drives the spread. A 90-day pet food cycle needs a higher first-order ROAS to be safe, because the channel is exposed to one extra quarter of churn risk before payback. A 30-day supplements cycle can tolerate a deeper first-order loss because the second purchase is almost in the same calendar month.
When negative first-order ROAS becomes a trap
There's a failure mode where teams keep funding sub-1x ROAS while the underlying cohort behavior quietly degrades. The dashboard looks the same week to week; the cohort never actually pays back. We cover the full pattern in churn-cliff diagnostics, but the short version: watch M2 retention, refill cadence drift, and welcome-offer abuse rate.
The other trap is welcome-offer depth. A 50%-off first order can drag first-order ROAS down to 0.4x and still pencil out if M2 retention holds — but the deeper the offer, the more discount-hunters you attract and the worse retention gets. Test the offer floor before assuming you can afford it.
Setting bids and budgets against the right ROAS
If you're bidding to first-order ROAS in Meta or Google, you're systematically under-spending on cohorts that retain well. The fix is to set bid caps and target ROAS at the LTV-adjusted number — typically your month-4 or month-6 cohort ROAS — and let first-order ROAS float as an output.
Practically, that means pushing a conversion value back to the ad platform that includes a probability-weighted refill expectation, not just the first-order revenue. Brands that make this switch usually find 20-40% of additional profitable spend they couldn't see before.
Frequently asked questions
On the first transaction, yes — you're spending €1.00 to acquire €0.80 of revenue, before COGS. But for a subscription brand, that's only the first slice of the customer relationship. If your M2 retention is 60% and your refill margin is healthy, cumulative gross profit usually crosses CAC by month 3 or 4.
Below roughly 0.5x you're typically funding too much discount, attracting too many one-and-done customers, or both. The exact floor depends on refill margin and M2 retention, but if you'd need M2 retention above 70% to make the math work, you're probably under the floor.
It's the same idea expressed differently. CAC payback period asks 'how many months until cumulative gross profit equals CAC?' First-order ROAS is just the month-1 slice of that same curve. The trap is treating month-1 ROAS as the whole story when it's only the first data point.
LTV-adjusted ROAS, almost always. Send back a conversion value that includes the probability-weighted refill expectation for that customer. Otherwise the algorithm under-bids on cohorts that retain well — exactly the customers you want.
Bring a cohort payback chart, not a campaign dashboard. Show last quarter's cohort cumulative gross margin crossing CAC on the timeline you forecast. Once they see the model and the actuals tracking each other, the conversation moves from 'cut the channel' to 'how do we scale it.'
Partially. If you have predictable repeat behavior — say, a coffee bean brand where 40% of customers reorder within 60 days even without a subscription — the same payback-curve logic applies, just with messier inputs. The further repeat behavior drifts from predictable, the more conservative you should be on first-order ROAS.
For most replenishment categories, refill 2 or 3. That maps to month 2-4 on a 30-day cadence or month 4-9 on a 60-90 day cadence. If you're not breaking even by refill 4, either your CAC is too high, your welcome offer is too deep, or M2 retention is structurally weak.
Usually 30-40% off is safe; 50%+ starts attracting discount-hunters whose M2 retention craters. The right way to find the floor is to A/B test offer depth and watch M2 retention by cohort — not just first-order conversion rate, which always improves with deeper discounts.
At minimum, one full refill cycle plus a 30-day buffer. For 30-day cycles that's month 2; for 60-90 day cycles it's month 3-4. Judging cohort health before refill 1 has happened is essentially judging on first-order ROAS — which we just argued against.
No — it's still a useful leading indicator and a guardrail against runaway spend. But treat it as one input, not the verdict. The verdict is the cohort payback curve crossing CAC on schedule.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.