# Trace survey pilot drop-off without inventing a reason

Reconcile stage counts and observed pilot evidence before attributing survey abandonment to a question.

If nothing was unclear or difficult, say so. Skip questions about steps you did not experience; use not applicable where appropriate.

## 1. What were you trying to do when you stopped or paused?

Answer: ____________________

## 2. Which instruction, if any, made the next step unclear?

Answer: ____________________

## 3. What other interruption affected this attempt?

Answer: ____________________

## 4. What would you need to decide whether to continue?

Answer: ____________________

## Author notes / illustrative keys where applicable

1. Ask about the actual stopping point without assuming dislike.
2. Locate a concrete instruction only if it affected this attempt.
3. Separate external interruptions from interface obstacles.
4. Collect a voluntary next-step need without pressuring completion.

Interpretation caution: Stage losses do not supply abandonment reasons. Only two unfinished cases have linked observed explanations; the other reasons remain unknown.


## Illustrative finding and follow-up

Observation: Ten starters minus six finishers leaves four unfinished attempts; only two have relevant observed transition notes.

Action to test: Clarify the transition and run another observed attempt; retain the unknown reasons instead of filling them in.

Follow-up: Ask new pilot volunteers to explain the transition while attempting it and inspect their stage log separately.


## Collection protocol

When and whom to ask: During an optional pilot of the actual draft, observe agreed attempts and distinguish pause, exit and finish.

Decision owner: The pilot facilitator records observations; the analyst reconciles distinct stage cases.

Follow-up: Ask new pilot volunteers to explain the transition while attempting it and inspect their stage log separately.

Response handling: Use synthetic counts here; explain observation and collection of timing separately. Allow people to stop.

## Decision workshop

Synthetic examples only. Manual worksheet; no automatic import or product writes.

### Synthetic pilot and reporting records

Stage or evidence | Count | Meaning
--- | --- | ---
Started | 10 | Known sequential cohort
Reached page 2 | 8 | 2 stopped before this stage
Finished | 6 | 2 more stopped before finish
Observed explanation | 2 | Linked to two of the four unfinished cases

### Version changes

Boundary | Draft | Revised
--- | --- | ---
Counts | Reason assumed from exit | Stage loss only
Observation | Two notes represent everyone | Two observed cases, other reasons unknown

### Draft or earlier version (review its limits)

Four people abandoned because the second page was confusing.

### Revised example

Four of ten starters did not finish in this invented single-path log. Two separate observed explanations mention the transition; the remaining reasons are unknown.

### Decision situations

- Stage identity and order are complete: Report the reconciled funnel, then investigate reasons separately.
- A relevant observed attempt is available: Use that explanation for its case without generalizing to all exits.
- Stage linkage is incomplete: Hold the funnel until distinct cases and branching are reconciled.

### Method references

- GOV.UK moderated usability testing: https://www.gov.uk/service-manual/user-research/using-moderated-usability-testing — Observed task attempts and participant interpretation; small pilots are not benchmarks.