breme.

Research · Measuring experience

Where subscription customers are really lost: what the churn data shows

By BREME · · 5 minute read

In short

Published churn data points to three quiet exits rather than one dramatic one. Customers stop using the product and cancel, payments fail and access lapses, and low price points attract customers who never commit. Recurly finds about a third of software churn is involuntary, and ChartMogul finds retention climbs sharply with price and annual plans. For customer success teams, each exit needs its own owner and its own early signal.

Who this is for: Customer success leads, heads of retention, operations managers and executives in software, SaaS and subscription businesses who need to know where renewals are actually lost.

The question: where does a subscription actually end?

Most churn reviews start with the cancellation and work backwards to a reason code. That makes churn look like a single decision made on a single day. The published data suggests something different: most subscriptions end through a few quiet exits that open well before anyone clicks cancel.

We looked at three sources that measure this from different angles. Recurly publishes churn benchmarks from its subscription billing network and an annual State of Subscriptions report. ChartMogul, a subscription analytics company, published a retention study of about 3,500 software companies in December 2025. Westpac commissioned a survey of 1,995 Australians about their subscriptions in July 2025.

None of them is a perfect picture of an Australian software business. Read together, they show where the customer success effort is most likely to pay off.

Exit one: customers stop using it

Recurly's 2026 State of Subscriptions, drawn from data on 76 million subscribers, reports that 52% of consumers cancelled at least one subscription in the past year because they were not using it. Lack of use is the headline reason, not price or service failure.

The Australian survey points the same way from the customer's side. Westpac found almost one in three Australians (32%) had paid for subscriptions they had forgotten about, and the most common cause of overspending was forgetting to cancel a trial before it auto renewed, at 38%.

A forgotten subscription is a customer who has already left in every way except billing. When they notice, they cancel, and they rarely come back to explain why. In software, the same pattern shows up as a customer who logs in less each month until the renewal conversation is a formality.

This is why we treat adoption as the leading measure and cancellation as the lagging one. By the time a cancellation arrives, the decision was made weeks or months earlier, during onboarding or the first quiet stretch after it.

Exit two: payments fail

Recurly separates voluntary churn, where a subscriber actively cancels, from involuntary churn, where a subscriber loses access because a payment fails. Its benchmark page, updated with July 2026 network data, puts SaaS churn at 3.22%, made up of 2.16 points voluntary and 1.06 points involuntary. By our calculation that makes roughly a third of software churn involuntary.

The share changes sharply with price. In Recurly's lowest tier of average revenue per customer, $10 to $25, involuntary churn is 1.30% of a 4.29% total. Above $250, it falls to 0.18% of a 3.07% total. Recurly notes that B2B businesses generally run lower churn than B2C because of longer contracts, higher switching costs and multiple seats.

Involuntary churn is the most fixable exit and often the least owned. A card expires, a retry fails, an email goes to a billing address nobody reads, and a customer who wanted to stay is gone. It usually sits between finance and customer success, and handovers nobody owns are where journeys break.

One caution on these figures. Recurly's own page reports slightly different totals for SaaS in different sections (3.04% and 3.22%), and mixes averages and medians. We use the proportions, not the precise rates.

Exit three: the customer never committed

ChartMogul's study, written by analyst Kyle Poyar, covered about 2,700 B2B SaaS companies, 600 B2C SaaS companies and 200 AI-native companies, limited to businesses with at least $250,000 in annual recurring revenue, using annualised 2025 rates.

Median net revenue retention for B2B SaaS was 82%, with the upper quartile at 97%. For B2C SaaS it was 49%. AI-native companies had gross revenue retention of 40% and net revenue retention of 48%, although their median gross retention rose from 27% in January to 40% in September 2025.

Price point mattered most. AI-native products priced above $250 a month kept 70% of revenue gross, which the report describes as essentially the same as B2B SaaS. Under $50 a month, gross retention fell to 23%. The report also finds that annual plans show median net revenue retention 10 to 20 percentage points higher than monthly plans.

The author's explanation is that products that are easy to buy are easy to cancel. Low priced, self serve buyers often see themselves as trialling, not committing, and paywalls can push them to pay before they have worked out how the product fits their business.

What the data does and does not show

These are vendor datasets. Recurly sees its own billing customers and ChartMogul sees companies it could classify from their websites, mostly outside Australia. Neither is a random sample of Australian software businesses, and the price figures are in US dollars.

They show association, not cause. Higher priced products may retain better because they are sold to more committed buyers, not because the price itself creates loyalty. ChartMogul's author says some comparisons rest on about 50 companies per group, so they are directional.

Westpac's survey is about consumer subscriptions, not business software. We include it because it shows the customer side of the same behaviour: subscriptions forgotten, trials left to renew, and one in three people finding it hard to cancel.

What we would put on the wall

If you run customer success, map the three exits as separate journeys, each with an owner and a signal that arrives early enough to act on.

  • Lack of use: define what active means for your product, and flag accounts whose usage drops for two cycles in a row, long before renewal.
  • Failed payments: decide who owns the moment a payment fails, check the retry and reminder sequence end to end, and make sure the message reaches a person who can fix it.
  • Weak commitment: look at retention by plan, price and contract length, and design onboarding for the customers most likely to treat the first month as a trial.
  • Give customers a way to pause. Recurly reports that among merchants offering pause before cancel, pause use rose 337% year on year, and three in four subscribers who pause eventually return.
  • Remember former customers. Recurly reports former subscribers drive nearly one in four new sign ups, so the way someone leaves shapes whether they come back.

The takeaway for customer success teams

Churn is rarely one bad moment. It is a set of quiet exits that open during onboarding, in the billing system and at the point of purchase, and each one is owned by a different team. The work is to make each exit visible early and give it someone who will act.

We are spending the next month on exactly these journeys for software and subscription teams. If that is where your renewals are leaking, start with breme.com.au/customer-success/.

Sources

See your journey from the customer's side

We map the end-to-end journey with you, find the moments that matter most, and build a strategy your teams and systems can deliver.

Start a conversation More insights