Subscribe & Save in 2026, when it’s still defensible economics for CPG brands

Subscribe & Save penetration is up across most CPG subcategories on Amazon. The headline reads as recurring-revenue strength. The math is more complicated, and the brands riding S&S into 2027 without re-running the unit economics are quietly underwriting a margin problem they’ll find in the Q4 board pack.

This post is the worked version of the audit we run on every CPG account: where S&S still pays, where it’s a trap, and the three numbers that decide which cohort you’re actually in.

The 2026 baseline

• S&S penetration in CPG categories: 39-52% of GMV (varies by subcategory)
• Standard discount: 5% on first delivery, 10-15% on five-or-more active subscriptions
• Coupon stacking: still allowed but tightened in Q1 2026, the legacy 20% S&S + 10% coupon combos are flagged more aggressively now
• Auto-renewal retention by month: ~85% month 1, ~64% month 3, ~38% month 12
• Median CAC for an S&S subscriber in CPG: 2.2x the equivalent one-time-purchase CAC, fully loaded

The 2.2x number is the one most operators don’t have. Amazon doesn’t expose it cleanly; you have to model it from search-term share, sponsored-placement-driven first-purchase data, and the lift attributable to S&S-eligible checkout placements. Brands that haven’t done the modeling assume S&S subscribers cost the same to acquire as one-time buyers. They cost more.

When S&S still pays

The math works cleanly under three conditions:

Condition 1: Replenishment cycle ≤ 60 days. S&S only earns its discount when the subscriber actually consumes the product on a predictable cadence. If your replenishment cycle is 90 days or more, the subscriber is more likely to skip, pause, or cancel before the second shipment. The 38% twelve-month retention number masks a wide spread: brands with 30-day cycles retain 50%+; brands with 90-day cycles retain 22%.

If your subscriber is canceling before the third shipment, you’ve paid 2.2x acquisition for one purchase plus a small fraction of a second. That’s not an S&S program. That’s a discounted first purchase with extra friction.

Condition 2: Margin floor ≥ 38% before discount. S&S costs you 10-15 points of contribution margin in steady state (the discount itself, plus the slightly higher pick-and-pack handling, plus the coupon stacking when it applies). If your gross margin pre-discount is below 38%, the S&S unit economics don’t work even at 90% retention. Most brands that keep pushing into S&S anyway are funding the discount out of marketing spend, which works for two quarters and then doesn’t.

Condition 3: Subscriber LTV : CAC ≥ 4:1 fully loaded. Including the higher S&S CAC, the discount drag, and the realistic retention curve. If you’re at 3:1, you’re treating LTV as the same as a steady-state subscriber when the curve actually decays. The brands that survive S&S as a percentage of their P&L into 2027 are the ones modeling at 4:1 or better.

When S&S becomes a trap

The trap pattern shows up most in pantry-adjacent and personal-care subcategories, where:

• Replenishment cycles run 60-120 days (so retention curves are punishing)
• Margins are 30-36% pre-discount (so S&S leaves single-digit contribution)
• S&S placement at checkout drives a meaningful share of conversions, so brands feel locked in
• Cancellation friction is light (Amazon’s UX optimizes for keeping customers, not subscribers)

The pattern: a brand notices that S&S is 40%+ of GMV, treats it as durable revenue, prices the rest of the catalog accordingly, and then runs into a quarter where retention dips 5 points and the contribution-margin number falls off a cliff. We’ve seen this happen six times in the last eighteen months across our CPG book; in every case, the brand thought they had a recurring-revenue moat and discovered they had a discounted-first-purchase machine.

What we audit on every CPG account

The S&S audit takes about 4 hours and surfaces whether the brand is in the works-cohort or the trap-cohort. Three pulls:

First, retention by cohort month. Amazon’s Brand Analytics gives you S&S subscriber count by month; the cancellation/skip rate has to be modeled from delivery-vs-active-subscriber gaps. We compute month-1, month-3, month-6, and month-12 retention separately. If month-3 is below 55%, the cycle isn’t matched to consumption, that’s the first trap signal.

Second, contribution margin pre- and post-S&S by SKU. Brands often run the discount math at the AOV level, which masks the SKUs where S&S is destroying margin. We isolate it per SKU and flag any below 22% contribution-margin-after-discount.

Third, S&S-driven CAC. We back into this from sponsored-placement spend allocated to S&S-checkout-eligible page views, divided by net new subscribers in the same window. Most brands haven’t run this calculation. The ones that have are usually surprised.

The 2026 economic shift

The reason this audit is more urgent in 2026 than it was in 2024 is that two things changed simultaneously: ad-driven CAC went up (median +18% YoY across our managed CPG accounts), and S&S retention went down (~3 points across the same population). Both compress the LTV : CAC ratio. The brands that were comfortably at 5:1 in 2023 are at 3.5:1 today and headed toward 3:1 unless they actively defend the curve.

Defending the curve has three pieces: improve retention (via product-led work, better packaging, clearer dosing, communication around the second-delivery moment), reduce CAC (via the Cosmo-aware A+ work that improves CTR and conversion on the same bids), and prune the SKUs where the math no longer works at all. Most brands resist the third piece. The brands compounding through 2026 are the ones doing all three.

What to do this quarter

If you haven’t run the audit, run it. It’s a single-quarter project that pays for itself within two billing cycles regardless of what it finds.

If your audit shows you’re in the works-cohort: keep going, but instrument the curve. Watch retention monthly, not quarterly. The catalog-as-product mindset matters here too, your S&S program is a product, and it benefits from versioned changes and a measurement loop.

If your audit shows you’re in the trap-cohort: stop pushing S&S as the primary growth lever. Reprice the SKUs where the math doesn’t work. Cut the discount on the trap SKUs and let those subscribers churn, the GMV loss looks bad in one quarter and reads as margin recovery in the next two.

The brands that compound on Amazon CPG over the next three years aren’t the ones with the largest S&S subscriber base. They’re the ones with the cleanest math underneath it.

Get an audit, we’ll model your S&S unit economics by SKU and tell you which cohort you’re actually in.