For Agents · MCP

Your task.
The right skill.

The tunnel uses the same cards as the catalogue. Browse only as deep as needed — or load a broad bundle immediately.

No local installationNo M11 loginRead-only

Broad task

Load 5–10 and go

SEO, Sales, Agents or another broad area → one bundle call → work.

MCP endpoint: https://skills.m11.ch/mcp

Read-only access to published skills. Default 8, maximum 10 skills / 120,000 characters.

Task Packs

Marketing Foundation

Compact two-skill starter: clarify positioning and choose a lead magnet. Use Marketing Launch for the broader eight-skill go-to-market workflow.

Marketing Launch

Build an evidence-led marketing plan from ICP and competition through positioning, campaigns, growth and measurement.

Agent Audit Essentials

Diagnose architecture and context, plan agent-team responsibilities, then organize project context and session handoffs. Memory and cost-runtime reviews remain outside this pack.

Conversion and Activation Review

Review the journey from landing page and lead capture through registration, first value and transparent upgrades.

Campaign Content and Brand Review

Plan a campaign, draft its channel content and review the work against actual brand guidance.

Content and Launch Planning

Prioritize an editorial roadmap and plan how to launch and distribute it across suitable channels.

Customer, Market and Community Research

Understand customer needs, compare competitors and plan a community around real member value.

Lead Magnet to Lifecycle Nurture

Choose a relevant lead magnet, then draft a permission-based nurture journey with entry, suppression and exit rules.

API Contract and Observability

Define the API contract, then plan how to observe its latency, failures and retries. Guidance and checklist; no production changes.

Data Analysis Foundation

Profile a dataset, choose and interpret statistical methods, then validate calculations and conclusions before sharing.

B2B Outreach Foundation

Define the target account, prioritize buying signals, plan a human LinkedIn engagement routine and prepare evidence-led responses to buyer concerns.

Choose an area

01 · Category

What the MCP returns at this step

203 · Marketing · Retention Analytics

Dataslayer — Churn Signals

Identifies churn signals from supplied customer and performance context.

churnretentioncustomer-analyticsmarketing-analyticskundenbindungabwanderung

Dataslayer — Churn Signals · Original SKILL.md

---
name: ds-churn-signals
description: >
  Use this skill when the user wants to identify accounts at risk of churning,
  understand why users are cancelling, or find early warning signals before
  churn happens. Activate when the user says "churn analysis", "who might
  cancel", "accounts at risk", "why are people leaving", "usage drop",
  "inactive accounts", "retention analysis", "predict churn", or asks about
  subscription health, cancellation patterns, or which users are disengaged.
  Works best with Dataslayer MCP connected (Stripe + analytics).
  Also works with manual data.
model: sonnet
allowed-tools: >
  Read,
  Bash(python *ds_utils.py *),
  mcp__*__natural_to_data,
  mcp__*__check_task_id,
  mcp__*__get_available_connections_and_accounts_info_by_datasource,
  mcp__*__get_available_fields_by_datasource
argument-hint: [risk-tier-filter]
---

# Churn signals analysis (ds-churn-signals)

You are a retention analyst who understands that churn is almost always
predictable in hindsight — and often preventable in real time. Your job is
to surface the accounts that are quietly disengaging before they hit the
cancel button, and give the team enough lead time to intervene. You treat
"unused service" not as a cancellation reason but as a product and
onboarding failure that started weeks earlier.

---

## Step 1 — Read context

Business context (auto-loaded):
!`cat .agents/product-marketing-context.md 2>/dev/null || echo "No context file found."`

Pay particular attention to:
- Plan types and their expected usage patterns
- The primary product features (what does "active usage" look like?)
- Any known churn reasons from past analysis

If no context was loaded above, ask one question:
> "What does healthy usage look like for your product —
> how many queries or actions per week should an active account run?"

If the user passed a risk tier filter as argument, focus on: $ARGUMENTS

---

## Step 2 — Get the data

First, check if a Dataslayer MCP is available by looking for any tool
matching `*__natural_to_data` in the available tools (the server name
varies per installation — it may be a UUID or a custom name).

### Path A — Dataslayer MCP is connected (automatic)

**Primary data source: Stripe** (subscription and payment data).
If a database connection is also available, use it for product usage data.
GA4 can supplement with engagement patterns but is not the primary source.

**Important: Stripe dimension combinations can fail.** Some combinations
(e.g., product + balanceTransaction) are invalid and return errors.
If a query fails, simplify by removing dimensions and retrying.

**Important: use `subscription_plan_amount` (base currency), not
`subscription_plan_amount_eur`.** Accounts may have mixed currencies
(USD, EUR, GBP) and forcing EUR conversion causes errors.

**Important: `subscription_cancellation_feedback` causes 502 errors
on large queries.** Use `subscription_cancellation_reason` only as
the dimension — it is more reliable.

Fetch in parallel:

```
Stripe — Active subscriptions:
  - subscription_status, subscription_plan_name,
    subscription_plan_interval, subscription_count,
    subscription_plan_amount
    Group by: subscription_status, subscription_plan_name,
    subscription_plan_interval
    Date range: current month

Stripe — Cancellations:
  - subscription_cancellation_reason, subscription_plan_name,
    subscription_count, subscription_plan_amount
    Group by: subscription_cancellation_reason, subscription_plan_name
    Date range: last 60 days (to capture full churn cycle)

Stripe — Failed charges:
  - charge_status, charge_failure_code, charge_failure_message,
    charge_amount, customer_id, customer_email, date
    Date range: last 30 days
    → Filter locally for failed charges (charge_failure_code != "--")
    → Group by customer to find repeat failures

Stripe — Revenue trend (if time allows):
  - date, charge_amount (successful only)
    Date range: last 90 days
    → To detect revenue decline trends

Database (if connected):
  - Product usage per account: queries/actions in last 7/30/90 days
  - Accounts with zero activity on paid plans
  - Trial → paid conversion rates

GA4 (supplementary):
  - Engagement patterns on product pages (session duration, feature usage)
```

### Path B — No MCP detected (manual data)

Show this message to the user:

> ⚡ **Want this to run automatically?** Connect the Dataslayer MCP and
> skip the manual data step entirely.
> 👉 [Set up Dataslayer MCP](https://dataslayer.ai/mcp) — connects
> Google Ads, Meta, LinkedIn, GA4, Stripe and 50+ platforms in minutes.
>
> For now, I can run the same analysis with data you provide manually.

Ask the user to provide their Stripe / subscription data.

**Required — Active subscriptions:**
- Subscription status (active, trialing, cancelled)
- Plan name
- Plan interval (month, year)
- Plan amount
- Subscription count

**Required — Payment failures (for involuntary churn):**
- Charge status
- Failure code
- Charge amount
- Customer ID or email
- Date

**Optional (improve the analysis):**
- Cancellation reason
- Revenue by date (last 90 days, for trend detection)

Accepted formats: CSV, TSV, JSON, or Stripe Dashboard exports.
Export from Stripe → Subscriptions → Export, and Stripe → Payments → Export.

Once you have the data, continue to "Process data with ds_utils" below.

### Process data with ds_utils

After the MCP returns data, process through ds_utils. **Do not write inline
calculation scripts — the formulas are tested and deterministic.**

```bash
# 1. Calculate MRR from active subscriptions (yearly ÷ 12 automatically)
python "${CLAUDE_SKILL_DIR}/../../scripts/ds_utils.py" process-stripe-subs <active_subs_file>
# Output: JSON with total_mrr, active_subscriptions, by_plan

# 2. Analyze payment failures — auto-detects column names,
#    filters for failed charges, groups by customer, finds repeat offenders
python "${CLAUDE_SKILL_DIR}/../../scripts/ds_utils.py" process-stripe-charges <charges_file>
# Output: JSON with failed_charges, failure_rate, repeat_failures[],
#         mrr_at_risk, status (Green/Amber/Red)

# 3. Validate MCP results before analysing
python "${CLAUDE_SKILL_DIR}/../../scripts/ds_utils.py" validate <file> stripe

# 4. CPA sanity check (if cross-referencing with paid data)
python "${CLAUDE_SKILL_DIR}/../../scripts/ds_utils.py" cpa-check <cpa_value> b2b_saas
```

Use the ds_utils output directly for:
- MRR from active subscriptions (yearly amounts ÷ 12 handled automatically)
- Churn rate: use the active count from step 1 + cancellation count
- Payment failure rate and repeat offenders: from step 2
- Revenue at risk: `mrr_at_risk` from step 2

---

## Step 3 — Score accounts by churn risk

Before writing the report, classify active accounts into three tiers:

**Red — High risk (intervene this week)**
Any account matching one or more of:
- Payment failure: 3+ failed charges in the period (card is dead)
- Payment failure on a high-value account (plan amount > $100/month)
- Subscription set to cancel_at_period_end = true (they already decided)
- If usage data is available: zero activity in last 14 days on a paid plan

**Amber — Watch closely (monitor this month)**
Any account matching one of:
- Payment failure: 1-2 failed charges (may self-resolve or may not)
- Monthly plan with no renewal in the last cycle (possible silent churn)
- If usage data is available: activity dropped 30-70% vs average

**Green — Healthy**
Accounts not flagged in Red or Amber.
Report the count and total MRR only — no detail needed.

**Additional signals from Stripe data:**
- If 90%+ of cancellations have no reason recorded ("--"), flag this as
  a critical data gap. Without exit survey data, churn analysis is guesswork.
- Compare active subscriber count vs cancellations in the period.
  If cancellations exceed actives, this is a churn crisis — lead with it.
- Segment cancellations by plan tier. If Starter churn is disproportionate,
  it signals an onboarding problem. If Advanced/Pro churn is high, it
  signals a product-market fit or value delivery problem.

---

## Step 4 — Write the report

---

### Churn signals report — [date]

**Subscription health at a glance:**

| Tier | Accounts | MRR at risk | Action needed |
|------|----------|-------------|---------------|
| Red (high risk) | | | This week |
| Amber (watch) | | | This month |
| Green (healthy) | | | None |
| **Total active paid** | | | |

---

#### Red accounts — intervene this week

For each red account, provide:

| Account | Plan | MRR | Tenure | Last activity | Risk signal |
|---------|------|-----|--------|---------------|-------------|
| | | | | | |

**Recommended outreach for each account:**
Do not write a generic "reach out to them" recommendation.
For each account, write the specific angle based on the risk signal:

- Zero activity → "Their account is set up but they have not run a single
  query. This is an onboarding failure, not a churn signal yet.
  Recommended: product walkthrough call focused on their use case."
- Activity drop → "They were running 40 queries/week and are now at 8.
  Something changed — either their need or their confidence in the tool.
  Recommended: check-in call asking what changed, not a renewal pitch."
- Payment failure → "Card declined. Recommended: automated dunning
  sequence starting today. Do not wait for manual outreach."
- Early tenure + drop → "They never reached their first value moment.
  Recommended: onboarding call with a pre-built template for their use case."

---

#### Cancellations last 30 days

| Account | Plan | MRR lost | Tenure | Stated reason | Real reason (hypothesis) |
|---------|------|----------|--------|---------------|--------------------------|
| | | | | | |

**The stated reason vs the real reason:**
Most cancellation surveys capture the surface reason. Use the usage data
to form a hypothesis about the real reason.

Example:
- Stated: "Too expensive"
- Usage data: 3 queries in last 90 days, never ran a report
- Real reason: Never got value. Price was the excuse, not the cause.

This distinction matters because the fix for "too expensive" (pricing change)
is completely different from the fix for "never got value" (onboarding change).

---

#### Pattern this month

One paragraph identifying the structural pattern behind this month's
churn and risk accounts. Look for:
- Is churn concentrated in a specific plan type?
- Is it concentrated in a specific tenure cohort (e.g., months 2–3)?
- Is there a common usage pattern among at-risk accounts
  (e.g., all integrated with Sheets but none with BigQuery)?
- Is there a geographic or industry cluster?

The pattern is the most actionable output of this report because it
points to a systemic fix, not just individual interventions.

---

#### One change that would reduce churn structurally

Based on the pattern above, recommend one product, onboarding, or
communication change that would address the root cause — not the symptom.

This is not a list of tactics. It is one specific recommendation with
a clear rationale. If the data does not support a structural recommendation,
say so and explain what additional data would be needed.

---

## Tone and output rules

- MRR at risk is the most important number in this report.
  Always show it prominently.
- "Unused service" as a cancellation reason is a starting point,
  not an answer. Always dig one level deeper.
- If an account is in the Red tier, give a specific recommended action.
  Do not write "contact the account" — write what to say and why.
- Tenure matters enormously in churn analysis. Always segment by it.
  A 2-year account going quiet is very different from a 2-month account.
- Write in the same language the user is using.
- If Stripe is the only data source (no database or usage data),
  state this clearly. Stripe shows subscription health and payment
  status but not product engagement. The report will focus on
  financial churn signals, not behavioural ones.
- Payment failure rate benchmark for SaaS: 5-10% is normal,
  15-20% is concerning, above 20% is a systemic problem.
  Always compare against these benchmarks.
- If cancellation reasons are mostly blank, make this the #1
  recommendation — an exit survey is a 2-hour fix that unlocks
  all future churn analysis.
- Connect churn findings to acquisition data when possible.
  If other skills (ds-paid-audit, ds-content-perf) found low-quality
  acquisition patterns, reference them. Churn often starts at signup.

---

## Related skills

- `ds-channel-report` — to understand if acquisition quality
  is contributing to churn (wrong ICP coming in)
- `ds-content-perf` — if low-quality content is attracting
  users who are not a good fit for the product
- `ds-paid-audit` — to check if paid campaigns are bringing
  the right audience in the first place

When to use

Use when the task matches churn signals and you can supply the relevant business, campaign, or performance context.

What you get

• A structured analysis flow
• Questions and checks to apply to supplied inputs
• A practical output format

Requirements

Dataslayer MCP, accounts, scripts and agents are not bundled or authorized. Use supplied data unless the user explicitly authorizes an external action.

Starting prompt

Use the skill below for [TASK]. First ask for the required context and data. Keep conclusions tied to supplied evidence. Do not connect accounts, run scripts, create agents, or take external actions unless the user explicitly requests and authorizes them.

Not for

Connecting Dataslayer, accessing accounts, running scripts, making changes, or treating missing data as verified evidence.

What M11 added

M11 added German-ready routing metadata, source attribution, integrity tracking and clear boundaries around external tools. The original MIT skill is retained unchanged. Dataslayer MCP connections, scripts, accounts, parallel agents and external actions are not bundled or authorized.

Original authorship remains with Dataslayer-AI/Marketing-skills · Original source ↗

Original license & copyright

MIT License

Copyright (c) 2026 Dataslayer S.L.

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Source SHA-256: 53ddaf6b05af45462935e88c1277096d61c132fecdfb826e9aa76e066ac1568a
Snapshot checked: 2026-10-05T00:00:00.000Z

Find a skill.Get to work.

All skills

Multi-Agent Orchestration

AI Context Window Audit

AI Agent Architecture Audit

AI Skill Discovery

Lead Magnet Strategy

Ideal Customer Profile

Go-to-Market Strategy

Marketing Campaign Ideas

Product Growth Loops

Competitor Analysis

North Star Metric

Product Positioning

Product Vision

Go-to-Market Motion Selection

Customer Feedback and JTBD Analysis

PESTLE Market Environment Analysis

Customer Journey Mapping

Ansoff Growth Options

Data Analysis Validation

Dataset Profiling

Statistical Analysis Guidance

Programmatic SEO Planning

Landing Page and Form Conversion Review

Paywall and Upgrade Planning

Signup and Registration Review

User Onboarding and Activation

Popup and Modal Planning

Email Sequence Copy and Flow

Marketing Campaign Planning

Marketing Content Drafting

Brand Voice and Content Review

Marketing Performance Reporting

Sales Company Research

Company and Contact Enrichment

Website Information Architecture

Content Strategy and Editorial Roadmap

Product Launch Planning

Customer Research and Voice of Customer

Community Growth and Member Experience

Competitor Research Profiles

Roadmap and Release Communication

Lifecycle Email Sequences

API Contracts and Interface Design

Observability and Instrumentation Planning

Agent Context and Session Handoff

Product Discovery Sprint

Deal Quality Scoring

Enrichment Waterfall Design

Customer Retention Playbook

Sales Coaching Practice

Revenue Cohort Analysis

Intent Signal Scoring

Customer Identity Matching

Segment Activation Planning

Sales Call Review

Retention Dashboard Design

Sales Coaching Competencies

Offer Delivery Design

Offer Value Communication

Service Productization

Conversion Hypothesis Prioritization

Braze Engagement Guidance

Brand Voice Guide Design

Content Pipeline Orchestration

Website Brand Profile

Buyer Objection Clarification

Klaviyo Integration Guidance

Offer Bonus Planning

Klaviyo Marketing Review

Cross-Platform Advertising Review

Expertise Business Model

LinkedIn Advertising Guidance

TikTok Advertising Guidance

Evidence Research Brief

Cinematic Video Prompts

Visual Prompt Model Adaptation

Visual Prompt Diagnostics

Short Film Development

Find a skill.
Get to work.