Subscription ecommerce
statistics and benchmarks.

The trouble with subscription statistics

Most subscription statistics are hand-me-downs. A number gets published once, a blog quotes it, twenty more blogs quote the blog, and within a year the figure is holding up a thousand strategy decks and nobody can say where it came from. Try to find the source of “the average subscription churn rate is 10%” and the trail goes cold in about four clicks.

This report works differently. Every number in it comes from subscription brands we manage or have audited ourselves. We pulled the raw order counts, applied one set of definitions to all of them, and measured everyone over the same stretch of time: the first eight months of 2026.

We did it because a number you cannot compare is a number you cannot fix. Telling a founder their brand lost 40% of its subscribers this year tells them almost nothing. Telling them that the typical brand in this data lost 55.8%, that the best-run lost 26.8%, and that the worst lost 81.4%, tells them exactly where they stand.

Where these numbers come from

Everything below is drawn from our own client accounts: real subscriber counts, real orders, real failed payments, recalculated by hand from the underlying counts. Nothing is survey data and nothing is quoted from anyone else.

A few rules kept the comparison honest:

• One window. Every brand is measured over the first eight months of 2026. Every rate on this page is a year-to-date figure, not a monthly one. If you benchmark yourself against it, use the same window.
• One set of definitions. Platforms count things differently, so we never take a headline percentage on trust. We recompute every rate from the raw counts underneath it, the same way for everyone, and check our figure against the platform’s own before using it.
• Anonymity. Everything is aggregated. No brand, client or account is identifiable, and no figure traces back to any single business.
• Honest churn. When a subscriber pauses, we count it as churn. Most dashboards do not, which is why our churn figures will read higher than the one your platform shows you. Most paused subscribers never come back; the ones who do get counted back in when they return.
• What the bands mean. For each metric we show the full range, the middle half, and the median. If your number sits past the 75th mark you are ahead of three quarters of the brands measured. For churn, being low is what winning looks like.

For the statistically minded: when this report says one number “predicts” another, we tested it, and we only use that word where the relationship is strong. Everyone else can read on without missing anything.

The definitions, in one place

• Revenue from subscriptions. Subscription order revenue as a share of all revenue.
• Orders that start a subscription. Of all orders placed, the share that opened a brand-new subscription at checkout. Recurring rebills do not count toward this, so a big existing base cannot flatter it.
• Revenue from subscribers. Revenue from customers who hold a subscription, including their one-off purchases.
• Subscriber churn. Cancellations plus pauses across the window, as a share of everyone active in it.
• First-attempt payment success. Scheduled charges that go through with no retry.
• Failed-payment recovery. Of the charges that fail and reach a resolution, the share rescued in the first recovery cycle. Charges still working through the retry sequence are left out of both halves rather than counted as losses.
• Order survival. Of the subscribers who start, the share still there at their third, fourth, fifth and sixth order.

Why we refuse to measure retention by the month

Nearly every retention benchmark you have seen reports how many customers are still around after one month, three months, six months. We think that number is close to meaningless, and our data shows why.

A subscription puts the same question to the customer every time it bills. A brand that ships weekly asks it about fifty times a year. A brand that ships quarterly asks four. On a calendar the weekly brand will always look worse, not because it keeps customers less well, but because it asks more often.

That is what we found. In our data, how often a brand bills tracks its annual churn closely: the fastest billers cluster at the high end of the churn table and the slowest at the low end. The pattern is strong rather than perfect, and and the exceptions matter: the single worst churn figure in the set belongs to a brand on a monthly cycle, and two brands billing on identical schedules sit more than twenty points apart. Billing rhythm is not the only thing driving churn. It is simply the thing a calendar-based benchmark cannot see, and it is large enough to swamp what you were trying to measure.

So we count retention a different way: by orders. Did the subscriber come back for a third order? A fifth? A sixth? Count like that and every brand faces the same test, because every order is a decision, whether it arrives weekly or quarterly. Counted like that, the gap between fast and slow billers narrows sharply, and some of the weekly brands that looked like disasters on the calendar turn out to be the best retainers in the whole dataset.

Every retention figure in this report is counted by order. When you read “survives to a sixth order,” that means six separate decisions to stay.

Same customer loyalty. Wildly different churn rate. That is why we count orders, not months.

The benchmarks

1. How much revenue comes from subscriptions

The fastest read on any brand: is this a subscription business, or a store with a subscribe button?

This is the widest range in the report, and it is not a smooth slope. Brands in this data are either subscription businesses to the bone, with three quarters or more of revenue recurring, or they are stores that happen to run a subscription program on the side. Almost nobody sits in the middle. The industry rows mostly restate that divide: supplements sell replenishment, so subscription is the business itself, while beauty brands attach a program to a storefront where most people still shop one order at a time.

One quieter observation from the two rows sitting nearly on top of each other: subscribers here buy their plan and very little else. The extra spending you might hope subscribers do around their subscription barely exists at any brand we measured. That wallet is still on the table.

2. How many orders start a subscription

Of every order a brand takes, how many open a brand-new subscription at the checkout? This number is set by the product page, the plan picker and the offer. Email has nothing to do with it.

Two settings, and very little ground between them.

Checkout capture does not spread out. It sits in two settings. In most stores about one order in ten opens a subscription. In stores built around the plan it is closer to one in two.

That difference is not something you close by testing a better widget. The higher side is running a different store: the plan is the default, the one-time purchase is the exception, and the page is built so that subscribing is the easiest thing to do. Getting there is a decision about how you sell.

Before you envy that side, read on. It costs them later.

Churn and payments

3. How many subscribers a brand loses in a year

Every cancellation and every pause, counted against everyone who was active during the window. Remember: your own dashboard leaves pauses out, so it flatters you compared with this table. Low is what good looks like.

Fast billers churn faster by construction. Compare against brands on your own rhythm.

The typical brand in this data loses a little more than half of its subscriber base in a year. The best-run loses about a quarter. The worst loses four out of every five subscribers it has.

The industry rows are not a leaderboard. The beauty brands here bill on long cycles and the food brands bill weekly, and as the previous section showed, fast billers rack up churn simply by asking the stay-or-go question more often. This table tells you what annual churn looks like in the wild. It does not tell you who is good at retention. Section six does that.

4. What happens when the payment fails

Before a subscriber can decide to stay, their card has to work. This is the leak almost nobody budgets for: charges that fail before retention ever gets a say, and the automated retry sequence that either rescues them or does not.

At the typical brand, one scheduled charge in seven fails on the first try. At the weakest, it is closer to one in five. What separates brands is what happens next. The best in this data rescue nearly two thirds of their failed charges. The worst rescue barely a quarter. Same failed card, more than double the save rate, and the entire difference lives in settings the customer never sees: when the retries fire, whether card details update automatically, what the reminder emails say and when.

Same failed card. More than double the rescues.

One result we did not expect. We checked which number in this report best predicts whether a brand keeps its subscribers deep into the relationship. It was not the discount, not the checkout design, not the size of the program. It was the two payment numbers, and the stronger of the pair is the plainest one: whether the card goes through on the first attempt. The brands whose scheduled charges simply work are, almost one for one, the brands whose subscribers are still there at a sixth order, with the save rate close behind. Every marketing measure in this report is weaker than both. Retention is won in the billing system at least as much as in the inbox, and the save rate is the half you can change this week.

5. The second order

Of the second orders these brands actually scheduled, this is how many were ever paid.

At the typical brand, half of all scheduled second orders never get paid. Not half of first-time buyers failing to subscribe. These are people who already subscribed, whose second box was already on the calendar, and one in two of those orders dies before the money moves. The biggest killer is cancellation before the charge date. No welcome flow, no winback campaign, no loyalty program ever gets a chance at revenue that never billed.

Skipping is worth watching separately. Most brands see almost none of it, but brands that put a prominent skip button in the customer portal run skip rates of 10 to 15% of scheduled orders. A skip is politer than a cancellation, and sometimes it genuinely saves the relationship. Often it is a cancellation with a delay on it. If you offer it, track it like churn, not like engagement.

Retention by order

6. How long subscribers actually stay

The share of subscribers still active at each order, counted the cadence-proof way. Every brand faces the same test here, whether it bills weekly or quarterly.

The typical brand walks about a third of its subscribers all the way to a sixth order. The best keep nearly half. The weakest keep one in ten. The shape of the curve is the same everywhere: the steep losses happen between the second and fourth orders, and a subscriber who makes it to the fourth rarely leaves before the sixth. Most of the loss happens early.

The churn table made the weekly food brands look like the worst retainers here. Counted by orders they are the best, keeping 46.3% of subscribers to a sixth order. The calendar was marking them down for billing often. The supplement brands, ahead on every revenue table above, keep 21.5% to order six.

Three conclusions

What we take from it

Key takeaways

• The typical brand loses 55.8% of its subscribers in a year, counting pauses as the churn they usually become; the best-run lose about a quarter, the worst lose four in five.
• Half of scheduled second orders never get paid at the typical brand, mostly through cancellation before the charge, which makes the stretch between first and second order the most valuable ground in the entire lifecycle.
• One scheduled charge in seven fails on the first attempt, and the best brands rescue nearly two thirds of failures while the worst rescue barely a quarter, on configuration alone.
• The payment numbers predicted long-term retention better than any marketing number we measured, which puts dunning settings ahead of winback copy on the priority list.
• Churn tracks billing frequency closely, so calendar-based retention comparisons are close to meaningless; count orders survived instead, where every brand faces the same test.
• Checkout capture comes in two camps, one in ten orders or one in two, with nobody in between, and the high camp keeps no more subscribers long-term: signing up and keeping are different skills.

See where your own numbers sit

These bands only become useful next to your own numbers. The LTV Calculator takes nine of them and shows what your subscribers could be worth over the next twelve months, and where the upside hides.

The framework behind this report is public: the LTV Parthenon lays out the three pillars and nine numbers every YOCTO account is run against. This report is those numbers measured across real accounts. If you want the same extraction run on your own account, that is what our subscription marketing team does.

Frequently asked questions

Cite this report

Free to cite and reference with attribution to YOCTO.

YOCTO, Subscription Ecommerce Statistics and Benchmarks 2026. https://yocto.agency/subscription-ecommerce-statistics/