# Tiered Churn Intervention Page: https://alidemirbas.com.tr/en/lab/journeys/churn-risk-escalation Churn risk rises → assess evidence → intervene by risk level ## Purpose When churn risk crosses a threshold, choose a response based on the strength of the signals and check whether usage recovers. Goal: Relationship Recovery & Intervention · Category: Retention & Subscriptions ## Steps - **Trigger**: Churn risk crosses the defined threshold - **Action**: Assess signal count and strength, segment / LTV, renewal timing and recent behaviour - **Decision**: What is the churn-risk level? - Low: Push: Remind the user of one relevant value or feature - Medium: Is the user in an active app session? - High: Is WhatsApp permitted? - **Action**: Push: Remind the user of one relevant value or feature - **Wait**: Wait 3 days - **Decision**: Did usage recover or risk fall? - Yes: Risk reduced / relationship recovered - No: Is the user in an active app session? - **Decision**: Is the user in an active app session? - Yes: In-app: Highlight the relevant feature in the current session - No: Email: Show personalized value the user may be missing - **Action**: In-app: Highlight the relevant feature in the current session - **Action**: Email: Show personalized value the user may be missing - **Wait**: Wait 3–5 days - **Decision**: Did usage recover or risk fall? - Yes: Risk reduced / relationship recovered - No: Is WhatsApp permitted? - **Action**: WhatsApp: Ask whether direct help would improve the experience - **Decision**: Is WhatsApp permitted? - Yes: WhatsApp: Ask whether direct help would improve the experience - No: Email: Ask whether direct help would improve the experience - **Action**: Email: Ask whether direct help would improve the experience - **Wait**: Wait 2–3 days - **Decision**: Was the problem resolved or risk reduced? - Yes: Risk reduced / relationship recovered - No: What is the main churn reason? - **Decision**: What is the main churn reason? - Price / value: Email: Offer a more suitable plan or package change - Usage difficulty: Email: Share a usage guide or onboarding help - Technical problem: Move to technical support - Payment problem: Move to payment support - Dissatisfaction / complaint: Move to Complaint Ownership and Closure - No clear reason: Email: Show personalized value, new features or a relevant guide - **Action**: Email: Offer a more suitable plan or package change - **Action**: Email: Share a usage guide or onboarding help - **Handoff**: Move to technical support - **Handoff**: Move to payment support - **Handoff**: Move to Complaint Ownership and Closure - **Action**: Email: Show personalized value, new features or a relevant guide - **Wait**: Wait 3–5 days - **Decision**: Did usage or purchase recover? - Yes: Risk reduced / relationship recovered - No: Eligible for a retention offer? - **Decision**: Eligible for a retention offer? - Only when a defined offer exists, no support or payment case is open, and no offer was sent in this run. - Yes: Email: Send one personalized retention offer - No: No recovery; move to passive churn monitoring - **Action**: Email: Send one personalized retention offer - **Wait**: Wait through the offer window - The window comes with the offer; stop as soon as usage or purchase recovers. - **Decision**: Did usage or purchase recover? - Yes: Risk reduced / relationship recovered - No: No recovery; move to passive churn monitoring - **Exit**: Risk reduced / relationship recovered - **Exit**: No recovery; move to passive churn monitoring ## Guardrails - Do not treat a weak signal as a confirmed cause of churn. - Stop the next intervention if usage recovers. - Support, payment or complaint issues move to their own journey; send no offer in the same run. ## Core rule Increase the level of intervention only when the risk evidence and observed cause support it. ## When to set it up Reach for this journey when churn reports show that customers whose risk score crosses the threshold rarely recover on their own, and every one of them currently gets the same response. Before launch, make sure the score can be read together with signal count, segment or LTV, renewal timing and recent behavior, that active-session status and push, email and WhatsApp consent sit on the customer record, and that an offer eligibility rule is defined. If the rise in risk comes from an open support, payment or complaint case, run the journey that resolves that case first; a churn message will not cover it up. ## How to set it up - **Trigger and entry**: Entry is the moment churn risk crosses the defined threshold; the first step assesses signal count and strength, segment or LTV, renewal timing and recent behavior. Customers who cross it are split at random between the journey group and a control group that receives no intervention. - **Who must not enter**: Scope is the customer relationship; no second run opens for the same rise in risk. A customer with an open support, payment or complaint case stays in that journey and receives no offer here, and a single weak signal is not on its own a reason to escalate. - **Timing and channels**: Low risk: a push with one relevant value, then 3 days. Medium risk, or no recovery yet: in-app during an active session, otherwise email, then 3 to 5 days. High risk, or still none: WhatsApp where permitted, otherwise email, offering direct help, then 2 to 3 days; a reason-based email follows with 3 to 5 days more, then, if eligible, one offer email through its offer window. - **Exit and handoff**: Usage recovered or risk reduced at any check ends the journey as risk reduced, and the next intervention is not sent. Technical problems move to technical support, payment problems to payment support, complaints to Complaint Ownership and Closure; a customer not eligible for the offer, or not recovered after it, moves to passive churn monitoring. - **Frequency and consent**: At most five touches on the longest path, each tier opening only when the previous one brought no recovery; push, email and WhatsApp each need their own consent. The one retention offer needs a defined offer, no open support or payment case and no earlier offer in this run. ## How to measure success **Incremental usage recovery rate vs control**: What matters is how many more customers return to usage after the intervention than in the control group; a falling risk score alone proves nothing. Count every customer assigned to either group at the threshold crossing as the denominator, and those whose usage or purchases recover and whose relationship continues within the same follow-up window after assignment as the numerator; the difference between the groups is the journey's effect. A delivered or answered message does not change the denominator. Also track: - **Risk reduction rate**: Assigned customers whose risk score falls back below the threshold, as a share of all assigned; report score changes separately from behavior, because a lower score is not recovery by itself. - **Recovery per tier**: Track recoveries after the push, the in-app or email step, the help question, the reason-based email and the offer separately; if the higher tiers add nothing, escalating intensity is not paying off. - **Handoff share and reason mix**: Watch the distribution of churn reasons and the share of customers handed to technical support, payment support or Complaint Ownership and Closure; a high handoff share means the problem belongs to those processes, not to churn messaging. - **Opt-outs and complaints**: At-risk customers are already sensitive, so watch push opt-outs, email unsubscribes and WhatsApp complaints per channel; if these numbers climb the intervention may be hastening churn. ## What to watch when reading the result - A journey group that does not return to usage more often than the control group tells you the intervention moved the score, not the behavior; more opens, replies or a lower score do not rescue it. Base the call on the usage recovery gap. - Treat the result as valid for this tier order, these waits and the risk threshold you defined. Credit the recovery of a customer handed to technical support, payment support or Complaint Ownership and Closure to that process, not to this journey as well; and do not report a reason label drawn from one weak signal as a confirmed cause of churn. - When few customers cross the threshold and recovery is rare, the gap can be chance. Before you start, enter the control group's natural recovery rate and the smallest gap you want to see into the [sample size calculator](https://alidemirbas.com.tr/en/calculators/sample-size-calculator); do not compare tiers until you have reached the number of customers it gives you. - If a later tier went out after usage had already recovered, count that as a setup fault in its own right; a help question or an offer sent to a recovered customer costs money and invites an opt-out. ## Frequently Asked Questions ### Should every high-risk customer receive a retention discount? First distinguish risk level and observed cause. This flow uses value reminders for lower risk and relevant help as signals strengthen. An offer is a later option for eligible users. A discount does not resolve technical or payment problems. ### Does a churn score prove the customer will leave? No. It is a risk signal. Assess it alongside recent behavior, signal strength, renewal timing and customer context. A higher score does not justify identical intensive treatment for everyone. Investigate missing or stale data too. ### What should happen when the churn reason is unknown? Do not invent a cause. Offer relevant value or useful guidance and observe the response. Route identified technical, payment or complaint issues to the right support flow, stopping outdated churn messages after handoff. ### Does a lower risk score prove the churn intervention worked? Not alone. Track actual usage recovery, purchases and continued customer relationships over a suitable window. Report score changes separately from behavior. If recovery fails, this example moves to passive monitoring after a bounded offer attempt.