Quarters·Churn Intelligence·2026 Edition

Churn isn't a moment.
It's a migration.

Consumers don't stop spending with you. They move to a competitor, one transaction at a time, and real spending data shows the warning arrives one to nine months before the customer is gone, if you know where to look.

01 · The Thesis

The month they disappear from your ledger is not when they left. It's when they finished leaving.

Traditional churn analysis measures the end of a relationship. The actual signal lives in the weeks and months before it.

Most companies track churn retrospectively. A customer stops transacting. Revenue drops. The system flags a churned account. Retention campaigns fire. Usually, it's too late.

The problem with this frame is not the campaigns. It's that you're reacting to a decision made months ago. Consumers don't stop spending in a category. They reallocate. They try a competitor. They start splitting their wallet. They realize they prefer the new option. They quietly complete the migration.

The decision to leave happens before you know it.
Quarters sees how spending moves across a whole category at once: not just what customers spend with one brand, but what they spend with its competitors, in the same window. That is the gap between watching churn happen and predicting it.
Where churn transitions appear in our data, by category
Shopping & Lifestyle
33%
Essentials
28%
Travel & Transport
14%
Food & Beverage
14%
Entertainment & Other
9%
02 · The Three Signals

Three ways a migration shows up in the data.

Some build over months, some happen in a single cycle. Each one surfaces in the category before it ever reaches your own reports.

SIGNAL 01

Competitive Activation

A new merchant appears in your category and takes a real share right away, not a tentative first try. When Lyft took 74% of an all-Uber ride wallet in its first month, the default had already changed.

SIGNAL 02

Share Erosion While Revenue Holds

Your absolute revenue looks stable or even up. But total spending in the category is rising, driven by a competitor eating your wallet share. A subscription can stay flat while the spending around it climbs many times higher, all of it going elsewhere. The dashboard looks fine. It isn't.

SIGNAL 03

The Clean Break

Not every switch leaves a runway. Some happen in a single cycle, at a renewal or a decision point, researched and settled entirely outside your view. One insurance policy ran flat for months, then simply didn't renew, and a competitor's policy took its place the same month. No trial, no overlap, no warning in your own data. The only tell is the competitor's first charge, the moment it lands in the category.

Case File 01/Travel & Transportation
UberLyft
61 days
Lead time
Nov '25
Signal first appeared
Jan '26
Uber's share hit zero

Six straight months of nothing but Uber, one hundred percent of the ride-sharing wallet. Lyft had never shown up in the category at all.

Then in November, Lyft appeared and took 74% of the month's ride spend. By December almost the entire wallet was Lyft, with Uber down to a single, final trip. By January, Uber's share was zero, and it stayed there: February, March, nothing.

Wallet Share: Ride-sharing Category (May 2025 – Jan 2026)
Uber (losing)
Lyft (gaining)
100% 67% 33% 0% May Jun Jul Aug Sep Oct Nov Dec Jan SIGNAL 100% 26% 0% 74% 100%
Monthly Wallet Share Breakdown
Field Note

Uber didn't lose this wallet in January. It left in November.

December looks like a contest, with both apps in the wallet. It wasn't. A wallet that lopsided is not indecision; it's the tail of a decision already made. The new default was set the moment Lyft took 74% of the wallet in month one. Uber had 61 days to respond, and the shift was visible in the category the whole time.

Case File 02/Wireless · Telecom
Fido MobilePublic Mobile
61 days
Lead time
Nov '25
Competitor appeared
100→0%
Fido's wallet share

This is the case that matters most for any subscription or recurring-revenue business. For six straight months, Fido held 100% of the wireless wallet. But the billing was anything but steady: it swung sharply from one month to the next, sometimes by more than half. Every spike is the kind of friction that sends people looking.

In November, Public Mobile answered. A new flat-rate plan activated while Fido's billing quietly wound down over the next two months and then stopped. Paying both carriers briefly isn't indecision, it's the fingerprint of a number being ported, with the new plan already live while the old carrier's final bills trail off.

61 days of warning.
That's how long from the first competitive signal to the wallet going flat, and the erratic billing had been signalling friction for months before even that. A proactive offer could have landed in the window. Instead, Public Mobile has held the wallet ever since.
Wallet Share: Wireless (Fido vs Public Mobile, May 2025 – Jan 2026)
Fido Mobile (losing)
Public Mobile (gaining)
100% 67% 33% 0% May Jun Jul Aug Sep Oct Nov Dec Jan SIGNAL 100% 53% 47% 0% 100%
Monthly Bill Volatility: Fido Mobile
Field Note

Billing volatility is a pre-churn signal hiding in plain sight

For six months the billing swung widely with no competitor in sight. Each spike is the kind of friction that pushes people to look around. Public Mobile's flat-rate pricing answered it, and once the new plan activated in November, the two-month overlap was just the old relationship winding down.

Case File 03/Software · AI Subscriptions
Claude AICursor AI
61 days
Lead time
Nov '25
Competitor appeared
100→4%
Claude's wallet share

This is the case that should worry every subscription business, because from the incumbent's side, absolutely nothing changed. For four months, Claude AI ran like clockwork, one steady subscription that was 100% of the AI-tooling wallet. The charge never spiked, never lapsed, never wavered.

In November, Cursor appeared, and it kept growing. Over the next two months the total spending on AI tools multiplied many times over, and nearly all of it went to the competitor. Claude still collected its flat subscription, but its share of the wallet had collapsed from 100% to 4%.

Nothing moved on the dashboard. The wallet went 100% → 4%.
This is what signal #2 looks like in real data. Claude's dashboard showed a perfectly retained subscription. The category data showed a wallet that had already moved, with the subscription surviving as a rounding error inside a budget the competitor now owns.
Wallet Share: AI Tools (Jul 2025 – Jan 2026)
Claude AI (losing)
Cursor AI (gaining)
100% 67% 33% 0% Jul '25 Aug Sep Oct Nov Dec Jan '26 SIGNAL 100% 27% 4% 73% 96%
Monthly Wallet Share Breakdown
Field Note

A retained subscription is not a retained wallet

Every churn model built on the incumbent's own data would score this as perfect: no lapse, no downgrade, no support tickets. The flat subscription keeps arriving. But the category budget multiplied many times over and the incumbent captured none of it. When the cancellation eventually comes, it will look sudden. It wasn't: the wallet left in November, 61 days before the share hit 4%.

Case File 04/Insurance
TD InsuranceSonnet
1 month
Whole switch, start to finish
Oct '25
Switched at renewal
9 mo
Sonnet held 100% since

For five months, TD Insurance held the whole policy, like clockwork. Perfectly regular, zero competitor activity. Then in October, at renewal, it simply stopped. The same month, a Sonnet policy activated instead. No trial, no overlap, no wind-down: the switch was researched, quoted, and completed entirely outside TD's field of view, between one renewal and the next. Sonnet has held the policy for the nine months since.

Gone between two renewals.
Insurance churn concentrates at renewal, and the shopping happens where the incumbent can't see it. TD's own data looked perfect right up until the switch. In the category data, the shift showed up right away, not a quarter later when the non-renewal surfaced in a cohort report.
Wallet Share: Insurance (TD vs Sonnet, May 2025 – Jul 2026)
TD Insurance (losing)
Sonnet (gaining)
100% 67% 33% 0% May '25 Jun Aug Sep Oct Dec Jan '26 Mar Apr May Jul SWITCH 100% 0% 100%
Field Note

You can't see every break coming. You can still get ahead of it.

A clean break leaves no trail, so trying to predict it from spending alone is a losing game. The play is timing. Quarters knows which categories turn over at renewal and when a decision point is approaching, so you can put a targeted offer in front of people well before that moment, giving them a reason to stay before they disappear without warning.

03 · The Window

How much warning you actually get.

Days between the first competitive signal and complete churn, for the migrations that left a runway. Some more examples:

Netflix → Amazon
Streaming
273 days
Fido Mobile → Public Mobile
Telecom
61
Uber → Lyft
Ride-sharing
61
Claude AI → Cursor AI
AI Subscriptions
61
Value Village → Talize
Thrift Retail
31
Shell → Esso
Fuel
30
090 days180 days273 days
04 · For Brands

Churn prediction is not a data science problem. It's a data access problem.

Every brand in this report had access to their own transaction data. None of them had access to their competitors'. That asymmetry is where churn lives.

The decision plays out entirely outside your view, in spending at other merchants that never appears in your own reports. Quarters sits across the whole category: how spending shifts between brands, month over month. When a competitor starts winning share, the change shows up early, not a quarter later in your P&L.

1

Activate retention in the window

When a switch leaves a runway, it can run one to nine months. That's a retention window, not a post-mortem, and targeted offers through Quarters reach those users during the evaluation phase. When a switch is abrupt instead, we still surface the competitor's first charge within days, not quarters.

2

See your category share, not just your revenue

Your revenue can look flat while your wallet share is collapsing. Quarterly reporting hides this. Real-time share-of-wallet data surfaces the trend before it registers in any traditional metric.

3

Partner with Quarters

Brands in the Quarters network gain visibility into the behavioral signals that precede competitive switching. We identify at-risk customers before they're lost, not after.

See the signal before the churn.

Quarters gives brands real-time behavioral intelligence on how consumers are allocating their spending across your category, including your competitors.

Partner with Quarters