Churn Risk Scorecard: 20 Warning Signs

Most churn is visible before it happens — in billing data, usage data, and support conversations your team already has. This scorecard turns that visibility into a monthly habit: twenty warning signs, scored, grouped by where they show up, each with the fix to run when you find it. Work through it once a month, score honestly, and act on the red flags in order. It pairs with the fuller retention audit worksheet when a score tells you to dig deeper.

Read this before the table: the points are a triage tool, not industry data. They are weighted by how close each signal usually sits to a cancellation and how quickly you can act on it — nothing here is a benchmark, and your total means nothing compared to another company's total. It means everything compared to your own total last month. If you want the underlying mechanics of why subscribers leave in the first place, read why subscribers churn alongside your first scoring session.

How to score

• Score what you can see. Each sign is scored for your subscriber base as a whole — or, if you have the data, for a specific cohort or at-risk segment. Use your billing platform, product analytics, and support inbox. No guessing: if you cannot observe a sign at all, score it 1 and add "get visibility" to your fix list, because an unmeasured signal is itself a risk.
• 0 = healthy. The sign is absent or rare, and a working process covers it.
• 1 = watch. The sign appears occasionally, or you cannot measure it reliably yet.
• 2 = act. The sign is common, rising, or has no owner and no process.
• Billing signs carry an extra point. A billing warning is weighted one point higher than the same score elsewhere, because a failed payment can end a subscription this month without the subscriber ever deciding anything. That weighting is a judgement call for triage, stated openly — adjust it if your business works differently.

The scorecard: billing signals (7 signs)

#	Warning sign	Where to see it	Run this fix
1	Failed payments sit untouched for more than 24 hours	Billing platform failed-payment list, sorted by age	Start the recovery sequence the same day: first email, first smart retry, and a card-update link in every message.
2	Hard declines are being retried like soft declines	Decline codes on recent failed charges	Stop retrying hard declines entirely. Route those subscribers to a card-update ask instead of another charge attempt.
3	Cards expiring in the next 60 days have had no notice	Card expiry report in your billing platform	Send a pre-expiry update request now, while the current card still works and the ask is a favour, not an emergency.
4	There is a backlog of old failed payments nobody owns	Failed payments older than 30 days	Work the backlog deliberately: recent soft declines first, hard declines to the update flow only, oldest accounts last.
5	Dunning emails go out without your branding or a working update page	Send yourself the live sequence and click every link	Brand the emails, use one recognisable update page, and test the full path on a real phone before the next billing run.
6	Annual renewals charge with no reminder beforehand	Renewal settings and your email calendar	Add reminders well before the charge, including a card-expiry check, so the largest payment of the year is never a surprise.
7	You cannot state last month's recovery rate	Your own reporting — or its absence	Start tracking recovered payments against total failures monthly, in the same sheet every month, so the rate is comparable over time.

The scorecard: engagement signals (7 signs)

#	Warning sign	Where to see it	Run this fix
8	Subscribers go quiet in their first two weeks	Product analytics: activity by days since signup	Fix onboarding before retention. Add one guided first win and a day-3 check-in for accounts with no activity.
9	Core-feature usage is falling for paying accounts	Usage reports for your one or two defining features	Trigger a human or email check-in when a paying account's core usage drops sharply, while there is still time to help.
10	A price or plan change is coming with no save plan	Your roadmap and pricing calendar	Prepare the announcement, the FAQ, and the save offers before the change — not in the week the cancellations start.
11	Long-tenured subscribers get the same treatment as new ones	Your lifecycle emails and offers, side by side	Give tenure a meaning: early access, a thank-you, or a check-in. Loyalty that is never acknowledged quietly expires.
12	Winback is a single email, or does not exist	Your cancellation follow-up automation	Run a spaced sequence matched to exit reasons — check-in, value reminder, what's new, one offer, and a respectful close.
13	Nobody reads cancellation reasons in aggregate	Exit survey data, if it is collected at all	Review the top three reasons monthly. If one reason leads, that reason gets the next fix — product, price page, or save offer.
14	Dormant subscribers are discovered only when their card fails	Cross-check: inactive accounts against next renewal dates	Re-engage dormant accounts before renewal, honestly. A subscriber reminded of value renews; a surprised one disputes and leaves.

The scorecard: support signals (6 signs)

#	Warning sign	Where to see it	Run this fix
15	The same complaint appears in tickets week after week	Support tags or a manual tally of recent tickets	Escalate the pattern, not the tickets. One recurring complaint is a product fix and a save-offer update waiting to happen.
16	Billing questions wait in the same queue as everything else	Support response times on payment-related tickets	Prioritise billing tickets. A subscriber asking about a charge is often a save in progress — a slow answer turns them into a cancellation.
17	Support cannot pause, discount, or escalate a save	Ask your team what they are actually allowed to offer	Give support a small, clear save toolkit with limits they can use without approval. A save that needs three sign-offs arrives too late.
18	Angry tickets close with no follow-up	Resolved tickets with negative sentiment, 30 days on	Add a follow-up check a few weeks after a bad experience. Accounts saved by support and then ignored churn at the next renewal instead.
19	Cancellations mention a competitor you never track	Exit reasons and support notes mentioning alternatives	Start a simple competitor note: who, why, what they offered. Patterns here should shape your roadmap conversations, not your panic.
20	No one owns churn as a number	Your meeting rhythm and reporting	Name an owner and review churn monthly alongside revenue, using one consistent calculation every month so comparisons actually mean something.

Reading your total

Your maximum is 47 points: 40 from the twenty signs, plus up to 7 weighting points if every billing sign scores 2. Use these bands as plain-language triage, not as grades against anyone else.

Total	Reading	What to do this month
0–10	Covered. The basics have owners and processes.	Fix the single highest-scoring sign and re-score next month. Watch for drift, not drama.
11–22	Exposed. Known gaps are quietly costing subscribers.	Work the 2-point signs in billing first, then engagement. One fix per week is a realistic pace.
23–34	Leaking. Several systems are missing at once.	Stop adding tactics. Fix billing recovery first — it is the only leak that cancels happy customers — then rebuild in group order.
35–47	Critical. Churn is being decided by neglect, not by your product.	Treat this as an operational priority. Score again in two weeks, not a month, until the total falls below the leaking band.

The monthly cadence

Score on the same day each month, with the same person scoring and one other person challenging any 0 that was also a 0 last month — healthy scores are the easiest to award lazily. Track the billing numbers behind signs 1 and 7 in a fixed format using the recovery KPI template, and calculate the headline churn figure the same way every month by following how to calculate churn rate. Keep every month's sheet. The trend is the value: a billing group that drops from 12 to 5 after you fixed your recovery sequence is proof the work landed, and a support group creeping up two points a month is a warning you would otherwise meet at renewal season. Bring the top three red flags to whoever owns product and support, with the fix from the table already attached, so the meeting decides priorities instead of debating definitions.

The bottom line

Twenty signs, one hour a month, fixes in priority order. The scorecard will not retain a single subscriber by itself — the fixes will, starting with the billing rows, because a customer lost to an unworked failed payment never even got to decide. Score honestly, especially the rows you cannot measure yet, and let the trend line tell you whether your retention work is working.

Frequently asked questions

How often should I run the churn risk scorecard?

Once a month is the right cadence for most subscription businesses. Score on the same day each month, keep every sheet, and compare against your own previous totals. If your score lands in the critical band, re-score after two weeks until it falls, because the fixes in that range should move the number quickly.

Are the scoring bands based on industry benchmarks?

No. The points and bands in this scorecard are a relative triage tool for your own business, weighted by how close each signal sits to a cancellation and how fast you can act on it. They are not industry data and should not be compared with another company's score. Their value is in your trend over time and in the order they give your fixes.

What if I cannot measure several of the signals?

Score each unmeasurable sign as 1 and write down the visibility you need as its fix. An unmeasured signal is a genuine risk, because churn building in a place you cannot see will only surface as cancellations. Getting the measurement in place is often the highest-value fix on the sheet.

Which group should I fix first if all three score badly?

Fix billing first. Engagement and support problems lose subscribers who were already drifting, but billing problems cancel subscribers who intended to stay, and they do it on a schedule you can predict. Once billing recovery is running with an owner and a sequence, work through engagement, then support.

Should I score my whole subscriber base or individual segments?

Start with the whole base to get your trend line established. Once the monthly habit exists, scoring a specific segment separately — new subscribers, annual plans, or a high-value tier — is often more revealing, because a healthy average can hide one segment in serious trouble.