What the Data Says About Amazon Subscribe & Save
When customers really subscribe, what the product page change did to sign-ups, how to measure churn properly, and when first-order vouchers pay off.
Subscribe & Save looks like the obvious way to build repeat revenue on Amazon. Sometimes it is. But after analysing subscription data in Amazon Marketing Cloud across several consumable brands, I've found most brands are measuring it wrongly, pushing it on products it doesn't suit, and missing the moment subscribers are actually won.
This guide covers what the data shows: when customers subscribe, what changed on the product page, how to measure churn properly, and when a first-order voucher is worth it.
The second order matters more than the subscription
Start with the bigger picture. On every Seller Central consumable I analysed, a customer who came back for a second order was worth 2.8 to 4.5 times as much as a one-time buyer.
A repeat buyer who then subscribed added only another 15% to 55% on top.
So Subscribe & Save is a useful extra, but the big lever is almost always getting the second purchase in the first place. If your repeat rate is weak, fix that before obsessing over subscription numbers.
Most subscribers sign up on their first order
Between 75% and 86% of subscribers signed up on their very first order. Between 88% and 94% had signed up by their second.
That has two big implications:
- The product page is where subscriptions are won. Anything that affects whether a new customer ticks Subscribe & Save on their first visit matters far more than any later nudge.
- Customers who've reordered by hand two or three times have usually decided against subscribing. They're a reorder audience, not a subscription audience. Retarget them with a reorder message instead.
The change that halved sign-ups on some products
Subscribe & Save used to be the pre-selected option on many product pages. Customers now have to choose it. In the US this was reported from mid-2025. In the UK data I analysed, the break came in April 2026.
First-order subscription rates fell on every Seller Central product I measured:
| Product type | Relative fall in first-order subscription rate |
|---|---|
| Daily personal-care consumable | about −40% |
| Adult daily tablet (mature product) | about −45% |
| Daily children's supplement | about −35% |
| Adult daily tablet (trial-heavy) | about −35% |
| Course supplement | about −34% |
| Daily powder supplement | about −29% |
It hit organic and ad-acquired customers alike. On one product I broke it down, and about 85% of the drop came from the page change itself, with the rest from a shift towards more ad-acquired customers.
Two things to watch:
- Growth can hide it. On two products, total subscriptions kept rising because new-customer numbers grew strongly, while the rate per new customer fell by a third. With flat traffic, the same drop would have shrunk the subscriber base.
- Don't compare across the change. Any Subscribe & Save figure, test or voucher result that spans the switch mixes two different worlds.
Is Subscribe & Save worth it for your product?
It isn't free money. The standing Subscribe & Save discount means a subscriber has to buy noticeably more often than a non-subscribing repeat buyer just to break even. With a 10% discount, that's roughly 22% more orders.
It worked well on daily-use products where subscribers genuinely used more. On a daily children's supplement, subscribers were worth about 55% more than non-subscribing repeat buyers, mostly because they actually got through more product.
It worked badly on a one-off occasion product. There, subscribers placed 17% fewer orders than non-subscribing repeat buyers. They subscribed, reached the end of the occasion and cancelled.
And sometimes the opportunity is bigger than the subscriber base. On one hygiene brand, frequent manual reorderers made up 49% of revenue, while Subscribe & Save was only 12%. The loyal core was already there, just not subscribed.
Ad-acquired customers also subscribed less. Customers who found the brand organically signed up at 1.15 to 1.83 times the rate of those who came in through ads.
Measure churn properly
Most brands get this wrong, and it changes the answer.
- Measure churn per subscription, not per customer. Customers held 1.23 to 1.32 subscriptions each. Using customer counts gave a misleadingly low figure. On one product, churn moved from a 7% "floor" to about 12% once it was measured properly.
- Expect churn to vary by product. Monthly churn ran from 6% to 22%:
| Product shape | Typical monthly churn | Typical subscriber life |
|---|---|---|
| Daily habit products | ~6–10% | 10–16 months |
| Daily children's supplement | ~12% | ~8 months |
| Trial or occasion-led products | ~15% | 6–7 months |
| Course products | ~22% | ~4.5 months |
On course products, high churn is often just people finishing the course, not unhappy customers.
- Track one-and-done subscriptions. Between 27% and 38% of subscriptions took only one delivery. About 30% is normal for a healthy consumable, and it's a useful baseline for judging any test.
- Only read mature data. Young subscriber groups look terrible, because they haven't had time to reorder yet.
In every case I looked at, the subscriber base was shrinking or stalling because too few people were signing up, not because more were leaving. Two simple numbers make this clear: the replacement level (active subscribers × monthly churn), which is how many new subscriptions you need each month just to stand still, and the long-run size of the base (new subscriptions per month ÷ monthly churn).
Are first-order Subscribe & Save vouchers worth it?
They can be, but the response isn't straight-line. On one product, the sign-up rate by voucher level looked like this:
| First-order voucher | Sign-up rate |
|---|---|
| None | ~8.5% |
| 15% | ~9.2% |
| 20% | ~12.6% |
| 25% | ~18.2% |
The 15% voucher barely moved anything and didn't cover its cost. The 25% voucher nearly doubled sign-ups. "Vouchers don't work" and "25% is too generous" were both wrong for that product. The lesson is that a shallow voucher showing no effect tells you about that tier, not about vouchers in general.
A few practical points:
- The real discount is bigger than it looks. The voucher stacks on top of the standard Subscribe & Save discount. A 25% voucher came through at about 31.6% in practice. Build your break-even on the real figure.
- Put the voucher where new customers enter. On one range, a voucher on the entry pack was used about 44 times as often as a deeper voucher on the loyalty pack.
- Check it actually ran. One voucher's budget ran out on day 22 of 31, and sign-ups dropped to zero the same day. On another brand, a voucher listed as live had stopped working a month earlier.
- Judge it on new subscribers, not monthly revenue. Seasonality can easily be bigger than any voucher effect.
- Watch subscriber quality. A 15% voucher was quality-neutral on one product: one-and-done was 26.8% without it and 27.3% with it. That doesn't prove deeper vouchers are safe, so keep an eye on one-and-done when you test.
The bottom line
Subscribe & Save is decided on the first order, it's worth less than the second order itself, and it's easy to measure wrongly. Check how many new customers subscribe each month, measure churn per subscription, compare like with like around the product page change, and test vouchers properly before deciding they don't work.
Want to see your own numbers? I'm running a fixed-price AMC Customer Value Snapshot for 5 founding clients, covering lifetime value, repeat buyers, reorder cycles and what else customers buy. Amazon's 5-year purchase data in AMC is free to query until 31 December.
Or start with the free Amazon LTV calculator to see what a second order is worth on your product.
Source for the US timing: Expandfi.
