# Friction Map — Cancellation as a Trust Moment

**Scope:** Direct web subscribers from intent to cancel through confirmed exit or voluntary alternative  
**Method:** Portfolio synthesis using hypothetical interviews, support themes, path analytics, and policy review  
**Evidence labels:** `Observed` = direct signal would be available; `Inferred` = plausible synthesis; `Unknown` = requires research

> This map deliberately separates evidence from confidence. Synthetic examples show how I would structure discovery; they are not presented as real customer research.

## Journey map

| Stage | Customer job | Friction | Evidence status | Risk if ignored | Product response | Measure |
|---|---|---|---|---|---|---|
| Recognize | Decide whether the subscription still fits | The customer cannot distinguish a temporary mismatch from a permanent one | Inferred | Premature churn or passive renewal | Frame options by reason, not by discount | Reason selection + downstream choice |
| Find | Locate cancellation without asking for help | Cancellation is visually subordinate to upgrade and support actions | Observed signal to validate via path review | Distrust, contacts, abandonment | First-class settings entry | Time to start; support-assisted starts |
| Understand | Know billing, access, and data consequences | Dates and data policy appear after commitment or in separate pages | Inferred | Fear, screenshots, support contact | Consequence summary before final action | Backtracks; policy-link opens; trust score |
| Decide | Compare cancellation with one relevant alternative | Generic offer walls create choice overload and suspicion | Inferred | Accidental acceptance; lower confidence | At most one reason-matched response | Offer comprehension; voluntary acceptance |
| Confirm | Complete cancellation once | Ambiguous labels and retry states obscure success | Observed signal to validate via support tags | Duplicate actions; billing disputes | Explicit label, idempotency, durable state | Completion; duplicate requests; errors |
| Verify | Confirm that renewal is off | Confirmation exists only in email or disappears on refresh | Unknown | Repeat attempts; contact volume | Persistent “Cancels on” account state | Repeat starts; confirmation views |
| Recover | Preserve a credible path back | Customers fear lost work or surprise restart | Inferred | Irreversible-feeling exit; low win-back | Clear retention window; reminder before restart | Export; reactivation; complaint rate |

## Friction severity model

Score each friction on three dimensions from 1–3:

- **Frequency:** how often eligible customers encounter it;
- **Consequence:** financial, emotional, or operational harm; and
- **Recoverability:** how hard the harm is to reverse.

`Severity = Frequency × Consequence × Recoverability`

This is a triage tool, not a substitute for evidence. A severe but uncertain friction becomes a research priority.

| Friction | F | C | R | Score | Action |
|---|---:|---:|---:|---:|---|
| Unclear cancellation status | 3 | 3 | 2 | 18 | Must solve in MVP |
| Billing/access ambiguity | 3 | 3 | 2 | 18 | Must solve in MVP |
| Generic offer wall | 2 | 2 | 2 | 8 | Replace with one relevant response |
| Hidden entry point | 2 | 3 | 2 | 12 | Validate and fix before experiment |
| Re-entry anxiety | 2 | 2 | 1 | 4 | Clarify; measure after launch |

## Emotional arc

```text
Resolve → suspicion → uncertainty → evaluation → relief
          ^             ^
          trust risk    choice-quality risk
```

The key insight is that emotional friction peaks before the final click. Optimizing button conversion alone misses the job.

## Questions that would change the solution

1. How many starts are true decisions versus attempts to understand pricing?
2. Which questions cause support contact before cancellation?
3. Do users who accept a pause return to meaningful use or merely delay churn?
4. Are privacy/trust reasons being misclassified under “other”?
5. Does cancellation friction disproportionately affect keyboard, screen-reader, or low-bandwidth users?

## Research plan

| Question | Method | Sample | Decision unlocked |
|---|---|---:|---|
| Can customers predict what happens after cancellation? | Moderated comprehension test | 8–10 | Copy and information hierarchy |
| Which moments create distrust? | Think-aloud journey interview | 10–12 | Entry, offer, and confirmation design |
| Are alternatives genuinely relevant? | Concept ranking by stated reason | 60+ survey responses | Eligibility rules |
| Does accepted pause produce value? | Behavioral cohort review | One billing cycle | Count pause as retained value or not |
| Does the flow work with assistive tech? | Accessibility test + audit | 5 users + specialist | Launch readiness |

## Anti-friction is not always good

Some friction protects the customer:

- a clear consequence review before an irreversible action;
- explicit consent before a pause that later restarts; and
- a confirmation step when the account has multiple members.

The goal is not zero friction. It is **remove confusion, preserve intentionality**.