Put the events in order
For C-104, distinguish the appointment change, customer comment, service response and later account event. Use the event date where known and retain the date the information was recorded.
A retrospective cancellation reason can help explain an outcome, but it is not evidence that the same information was available for earlier intervention. Keep retrospective and contemporaneous observations distinguishable.
Build a comparison that fits the question
Compare customers at relevant service stages and within a defined observation window. Accounts that just joined have not had the same opportunity to reach a six-month outcome as older accounts.
Include retained accounts, cancelled accounts and accounts without completed check-ins. Report the usable evidence and exclusions. Otherwise, a dashboard can make respondents look like the full customer base.
Keep separate service types or account relationships where they affect interpretation. The same person may be a user, family contact or purchaser; their perspective and relationship to the account should remain visible.
Read associations carefully
If reported device issues occur more often among cancelled accounts, that identifies a question to investigate. It does not establish that the issue caused cancellation or that every account with the issue will leave.
Record follow-up actions and whether later feedback supports resolution. Customers receiving intervention may differ from others in ways that affect outcomes, so a simple comparison does not prove the intervention worked.
Use the connected record to ask better questions and improve service decisions. Preserve the sources, time window and limitations so the team can challenge a conclusion rather than accepting an opaque risk label.
Work through the Clearway example
This is a fictional teaching example. Adapt the fields and rules to the question your own workflow needs to answer.
| Check | Why it matters | Record |
|---|---|---|
| Sequence | Avoid using later information as an earlier signal | Event and entry dates |
| Coverage | Avoid describing respondents as everyone | Eligible, invited and responding accounts |
| Comparison | Different stages create different exposure | Service stage and outcome window |
| Action | Understand what occurred after feedback | Owner, response and confirmation |
Build this part of your plan in more detail
These lessons address the next practical questions in this module. Choose the detail your workflow needs, then return to complete the exercise.
- How to Change Survey Questions Without Losing Comparability
Check whether a changed survey still supports the same trend.
- How to Analyze Longitudinal Survey Data
Use actual observation dates when connecting customer feedback over time.
More practical questions in this module (1)
- How to Build a Relationship Record
Prepare a source-linked relationship handoff without confusing people, organizations and interactions.
Add this part to your plan
On workbook page 4, place C-104’s events on a timeline and define the accounts included in a comparison.
Download workbook (fillable PDF)Compare with a suggested answer
The timeline separates the original concern from later explanation. The comparison includes retained accounts and missing responses, states the observation window and does not claim causation.
Self-check: Can you explain which information was available before the outcome?
Questions you may have
Can we analyze cancellations alone?
You can describe those records, but evaluating associations with cancellation also requires an appropriate comparison population.
Does linked history prove why customers leave?
No. It improves the evidence available; interpretation still needs a suitable design and limits.
Apply the method to your work
Use the five-part plan to assess the sources, analysis and permissions your team needs. The solution page shows where a connected platform can support that workflow.
Explore the customer experience solution →