Most Amazon brands buy their sales every month. Ads bring in the customer, the customer buys once, and next month the ads have to do it again. The spend never falls, because nothing the spend produced stays.
Josh Grillo, who runs brand management here, puts the alternative simply: the most successful brands don't purchase all their sales. A brand compounds when the customers it won last quarter keep buying without being bought again, so ad spend can fall as a share of sales while revenue grows. That is the flywheel, and it has a failure mode that few dashboards show.
What the flywheel is
A flywheel is a heavy wheel that takes effort to start and then stores momentum. On Amazon, the effort is paid acquisition: ads, launch pricing, reviews. The momentum comes from three things that build on each other.
- Repeat purchase. A customer who buys again costs far less than a new one, because nobody has to be introduced to the brand a second time.
- Subscribe & Save. A recurring order turns a purchase into a commitment and removes the decision each month.
- Organic rank. Steady unit velocity pushes the listing up in search, where it gets seen without an ad.
When all three work, a growing share of the sales are earned. The paid cost of each new-to-brand customer falls away because the organic base is carrying more of the load.
The failure mode: acquisition slower than attrition
The flywheel runs backward when a brand loses customers faster than it wins them. This is the number most brands don't measure.
We often see it on audits of subscription-heavy accounts. New-to-brand customers arrive every week, which sounds respectable until you set them against the other side, where more customers are lapsing than arriving. Amazon's Customer Loyalty Analytics, inside Brand Analytics, groups your buyers by how recently and how often they've bought, and that's where to look. The account looks stable because revenue is flat, and revenue is flat because price increases and larger packs are covering for falling units.
Some illustrative arithmetic shows how fast this works. Say a brand has 1,000 active customers, wins 60 new ones a month and loses 8% of its base each month. That's 80 gone and 60 arrived, so the base shrinks by 20. It keeps shrinking until it reaches 750, because that's where 8% of the base equals the 60 coming in. A customer base settles at new customers per month divided by the monthly lapse rate.
That one division changes the plan. To hold 1,000 customers at an 8% lapse rate, the brand needs 80 new customers a month, and every one of them has to be bought. Bring the lapse rate down to 5% and the same 60 new customers support a base of 1,200, without another dollar on acquisition.
A brand in that state is paying to acquire customers and losing them as fast as they arrive. Running more ads only raises the cost, so the first job is to stop the leak.
How does Amazon Subscribe & Save work for sellers?
Subscribe & Save lets a shopper schedule recurring deliveries of a product at a discount, and the seller decides whether to fund that discount at 0%, 5% or 10%. Amazon's seller program page, checked in November 2026, says many eligible Fulfillment by Amazon products are enrolled automatically at a 0% seller-funded discount, so a product can be in the program without anyone on your team having chosen it.
Amazon funds a separate 5% when a customer has five or more Subscribe & Save items in one delivery. Its seller guide says Amazon-funded discounts are added at Amazon's discretion, so treat that tier as Amazon's money and Amazon's decision, and don't build a plan around it.
| Your base funding | What the shopper gets on four items or fewer | With five or more items in the delivery |
|---|---|---|
| 0% | No discount | 5%, funded by Amazon |
| 5% | 5%, funded by you | Up to 10% |
| 10% | 10%, funded by you | Up to 15% |
Amazon says there are no additional selling costs for the program, and that participants pay standard selling fees and normal Fulfillment by Amazon costs. The discount is the cost, and it comes out of your margin on every delivery. You manage funding, enrollment and metrics on the Subscribe & Save page in Seller Central.
The first thing to do is look at that page. Because enrollment can be automatic, a brand can have products sitting at 0%, where a shopper ordering fewer than five items sees no reason to subscribe.
Which products qualify for Subscribe & Save?
Amazon decides eligibility by category, average selling price, sales performance, fulfillment history and in-stock rate, and the product has to belong to a brand enrolled in Brand Registry. You also need to be a Brand Representative for that brand and have a buyable offer.
Amazon's own guide for sellers, published in August 2025, puts the in-stock requirement at above 90%. Many third-party guides still quote 85%, so go with Amazon's number and check the eligibility tool in your own account, because Amazon doesn't publish a full rule set outside Seller Central.
Fulfillment by Amazon is the simple route, and products have to be sold through it to be auto-enrolled. Seller-fulfilled products can get in, but Amazon's guidance has more conditions: its guide says to request each product and lists free domestic shipping and delivery estimate accuracy above 93.5%, while a 2024 forum announcement describes automatic enrollment for eligible seller-fulfilled products. If you ship your own, confirm what your account shows before you plan around it. FBA versus FBM covers the wider choice.
For a food brand, consumables are the best candidates. Pantry staples, snacks and supplements are used up and bought again, which is what a subscription needs. If your product is a one-time purchase, the flywheel runs on referrals and reviews instead.
How steadily the product gets used matters as much as the category. Velocity Sellers, an agency that publishes figures from its own portfolio of more than 140 brands, reports cancellation under 25% per delivery cycle on predictable consumables and above 60% on bulk goods that shoppers stockpile. It doesn't publish its method, so read those as directional. They match what you'd expect: someone who drinks one coffee a day needs the next bag, and someone sitting on a year of paper towels cancels.
That's worth holding against your pack decision. A larger pack can suit a subscription, but it also raises the price per delivery and lengthens the gap between orders. Josh's view is that pricing under $20 lifts both conversion and subscriptions, for the same reason people forget about a streaming charge on a credit card: below that level, most people don't think it's worth their time to cancel. That's his read and a pattern to test, and it's a reason to run a subscription on more than one size.
What discount should you set on Subscribe & Save?
Set the highest base discount that still leaves the SKU above your margin floor, and for a consumable that's the 30% floor with the discount included. A subscription priced below your floor makes the flywheel spin the wrong way.
Here's an illustrative SKU, with made-up numbers. A $30 bag has $18 of true landed COGS and Amazon fees, leaving $12, or a 40% pre-ad gross margin. Pull your own fees for this, because they change by category and price point. We've held fees flat in this example, which slightly overstates the cost of the discount, since a fee charged as a percentage of price falls with the price.
- At a 5% discount, the bag sells for $28.50, keeps $10.50 and runs a 36.8% margin.
- At a 10% discount, it sells for $27, keeps $9 and runs a 33.3% margin.
Worked backward from the 30% floor, that gives a simple rule. A SKU needs at least a 33.5% margin before the discount to carry 5%, and at least 37% to carry 10%. Under 33.5%, stay at 0% and fix the price, pack or COGS first.
Amazon's case for going deeper is its claim of up to a 1.8x increase in conversion for products with a 10% to 15% discount, sourced to Amazon internal data. That's an upper bound from the company that runs the program, and it says nothing about what the sale kept. Test it on your SKU and judge it by profit per delivery.
Put the discount on its own line. Coupons and subscription discounts belong in the promotions row of your P&L, and stacking them with other offers is how a sensible price ends up selling at a loss. The margin piece has an example of four reasonable decisions adding up to that.
How many extra orders does a subscriber need to pay back the discount?
A subscriber pays back the discount once they place more orders than they would have placed anyway, by a multiple equal to your margin divided by your margin minus the discount. At a 40% margin and a 10% discount, that's 40 divided by 30, or 1.33 times as many orders.
This is the question most Subscribe & Save guides skip. A discount given to a shopper who would have reordered anyway at full price is money you've handed back, so the subscription has to create extra orders to earn it.
Back to the illustrative bag. A repeat customer who buys three times a year at full price brings you $36 of pre-ad margin. As a subscriber at 10% off, the same person needs four deliveries to match that, at $9 each. Five deliveries puts you ahead, and three leaves you $9 worse off than if they'd never subscribed.
| Margin before discount | Margin after 5% off | Orders needed vs. no subscription | Margin after 10% off | Orders needed vs. no subscription |
|---|---|---|---|---|
| 30% | 26.3% | 1.20x | 22.2% | 1.50x |
| 35% | 31.6% | 1.17x | 27.8% | 1.40x |
| 40% | 36.8% | 1.14x | 33.3% | 1.33x |
| 50% | 47.4% | 1.11x | 44.4% | 1.25x |
The table is arithmetic with fees held flat, and your own numbers will differ. Thin margins are punished twice, because the discount takes a larger share of what's left and the subscriber has to stay longer to earn it back.
Compare the multiple with how long your subscribers stay. Velocity Sellers puts the average for consumer packaged goods and supplements at 2.4 to 3.8 deliveries before cancellation, again from its own data. If a subscriber stays for three deliveries and would have bought twice anyway, a 10% discount at a 40% margin clears the 1.33x bar. If they'd have bought three times anyway, it doesn't.
What is a subscriber worth?
A subscriber is worth what they keep ordering at their discounted margin, minus what it cost to win them, and in the accounts we review subscribers are often worth several times what a one-time buyer is. That's why the subscription base is frequently the most valuable asset in a consumables account, and why a falling subscriber count deserves more attention than a flat revenue line.
Pull the number for your own brand from the Subscribe & Save dashboard in Seller Central. Amazon's help page says it covers reorder performance. Reports from agencies describe a view of lifetime value and reorder rate split by subscribers and non-subscribers, along with out-of-stock losses, but what you see can vary by account. Amazon's guide also lists the share of sales from Subscribe & Save, active subscribers and retention count. Check the month-over-month trend, because the total alone hides the direction.
When a subscription base shrinks, the reflex is to raise the offer. We often see a large coupon stacked on the standing discount that still doesn't pull people in. When a discount that deep fails, look at the foundation: the listing doesn't make the case for the product, the reviews don't support trust or the reason to reorder isn't clear. Josh describes the same thing with his coin analogy, where ads bring qualified traffic and the listing converts it. Traffic sent to a listing that doesn't convert starves the flywheel of its subscribers and repeat orders.
Should you use a Subscribe & Save coupon?
Use a coupon when you want more sign-ups without lowering what every later delivery keeps, because Amazon applies it to the first delivery of a new subscription while the base discount runs on all of them. Amazon's seller guide says these coupons can be a dollar or percentage amount and can run up to three months.
On the illustrative bag, a 15% coupon on top of a 10% base brings the first delivery to about $22.50, and that delivery keeps about $4.50, a 20% margin before any coupon fees. Check how your own offer combines in the coupon preview, since we've simply added the two here. The next deliveries go back to $9.
So the coupon costs $4.50 once, and a deeper base discount costs you on every delivery for as long as the subscriber stays. Going from a 5% base to a 10% base costs $1.50 per delivery. Across a three-delivery subscriber that matches the coupon, and across a longer one the coupon is cheaper.
Two conditions apply. A coupon brings in people who may cancel after the first delivery, and you'll see that in your retention count in the months after you run one. It also can't fix a listing that isn't converting.
Making the target countable
Say holding the first page for your main term takes 40 units a day. That figure is illustrative and yours will differ, but the exercise is worth doing for any brand. It turns "build rank" into a number you can manage to.
Work it backward:
- Estimate the units per day the top few competitors sell on your main term. Third-party tools give estimates, and they're ballpark. Brand Analytics shows where the clicks and purchases on that term go.
- Subtract your current organic daily units to get the gap.
- Split the gap between new customers you need to win and repeat orders you can earn.
- Set a monthly target for each.
Repeat orders are cheaper per unit, so the plan should lean on them where the product allows.
Subscribers turn into units with one more division: active subscribers divided by the days between deliveries. In an illustrative case, 450 subscribers on a 45-day cadence ship 10 units a day. If your target is 40 units a day and organic sells 15, that base covers 10 of the 25-unit gap, and it's 10 units you aren't paying an ad to find each morning. We haven't found Amazon saying how much a subscription order counts toward rank against any other order, so use this to size demand, and treat the rank effect as something you watch.
What happens to subscribers when you run out of stock?
A stockout costs you the deliveries that don't ship, and it gives subscribers a reason to leave. Keeping products in stock is, in Amazon's words, a key consideration for the program, and the in-stock rate is part of how products qualify.
Amazon's help page says that if a Subscribe & Save offer is unavailable for more than 7 days, it transfers the subscriptions to an identical, available offer from the same seller. If you don't have one, we'd plan on losing those deliveries. Amazon's performance dashboard has reported the percentage of units not delivered because the ASIN was out of stock, so check that line before you read a dip in subscribers as a pricing issue.
Forecast subscription demand yourself and hold it apart from one-time demand. Our advice is to rely on Amazon's tools as little as you can here, because retention comes from being consistent on inventory.
The same care applies when you clean up a catalog. Consolidating ASINs can strand subscribers on a listing you're about to retire. Amazon's help page says subscriptions can be moved to a new SKU under the same ASIN on the Manage Products page, and that transfers across ASINs go through your account manager or the Get help page. Do that before the old listing goes dark.
Three levers that feed it
Make the second order easy
Put the reason to reorder on the page. A bullet or an A+ module that says how long a bag lasts and what the next order saves does more than a deeper discount. Check that Subscribe & Save is enabled on the right SKUs and set at a discount that leaves your margin intact.
Reach lapsed customers with a lower-cost audience
Brand Analytics can show repeat purchase behavior, and Amazon's tools let you reach past buyers in limited ways. Brand Tailored Promotions, for Brand Registry brands, offers 10% to 50% off to groups that Amazon's guide describes as including repeat purchasers, loyal customers whose sales are declining and lapsed customers. Sponsored Display can remarket to people who bought your products, with a lookback window you can set to match how long the product lasts. Check what your account can see, because the availability of these reports has changed. Remarketing to people who already bought is cheaper than buying a stranger's first click, and it should sit at the low end of your acquisition cost.
Protect the reviews
Subscribers and repeat buyers write the reviews that new shoppers read. A product with a steady stream of recent reviews holds conversion and rank. The Vine program can seed a launch, and the long-run source is customers who keep ordering.
The measurements to watch
| Measure | What it tells you |
|---|---|
| New-to-brand customers per month | Acquisition |
| Customers lapsing per month (bought before, not in the last window) | Attrition |
| Active Subscribe & Save count, month over month | Commitment |
| Repeat purchase rate on consumable SKUs | Reorder behavior |
| TACoS over time | Whether paid share is falling as organic grows |
If acquisition exceeds attrition and TACoS is trending down while sales rise, the wheel is turning. If TACoS is flat or rising while units fall, ads are being used to replace customers who left. The TACoS piece has the test.
When to reinvest, and when to pull back
The pattern we aim for is to reinvest into ads while the organic base is still being built, and to ease off once rank holds and repeat orders carry a growing share. Brands that expect a year of losses before profit are often locked in an endless chase. A launch that is positioned correctly from the start, on a product with healthy margin, can pull back its spend and still make money. On a natural snack brand, we layered Subscribe & Save across the catalog as part of a wider rebuild of listings, ads and reviews.
This is also where the order of operations matters. Listing, then ads, then reinvestment is the sequence, and the flywheel is the third step.
The bottom line
A brand that buys every sale has to keep buying them. A brand whose customers return builds an asset, and the ad budget can fall as the base grows. Measure acquisition against attrition before spending more, and aim your plan at the second order.
If you want to know which way your wheel is turning, send us your ASINs and we'll look at your new-to-brand trend, your Subscribe & Save count and the margin each subscription keeps, and tell you what we see.






