Partial response

A respondent answered half the questions and closed the tab without clicking "Submit". Such answers are called partial responses: the data exists, but the session was never completed. In SurveyNinja, incomplete responses can be saved and reviewed — this is a platform feature. It helps you understand at which question respondents most often abandon the survey, estimate the dropout rate and the completion rate, and improve the survey by reducing drop-off.

Without saving partial data we only see those who reached the end. Those who left midway "disappear" from the statistics — and we have no idea at which question they stopped. Saving partial responses gives the full funnel picture and concrete points to fix.

What a partial response is in plain terms

Partial response — answers a respondent entered in a survey when the session was not completed by clicking the submit button: the respondent closed the page, interrupted the run, or never reached the final screen. In survey collection systems, partial responses may be saved (as in SurveyNinja) or discarded. Saving partial responses lets you analyze drop-off by question, calculate completion and dropout metrics more accurately, and find the "bottlenecks" in a survey.

Put simply: a partial response is "started, but never clicked Submit". If the platform saves it, we can see how far a respondent got and which answers they managed to give.

Not every platform saves partial responses by default; in SurveyNinja this option can be enabled in the settings, after which incomplete responses become available in the reports section alongside completed ones.

Why save partial responses

Drop-off analysis by question. Saved partial responses show at which question or screen respondents most often interrupt the run. This gives concrete points for simplifying wording, shortening a block, or changing the logic.

Accurate completion and dropout rates. To calculate the completion rate and the abandonment rate you need to know the number of "starters" — those who answered at least the first question. Saving partial responses lets you count starters correctly and avoid losing them in the statistics.

Understanding the reasons for drop-off. Sometimes partial responses reveal that respondents "drop out" at the same question (for example, a complex or personal one). This tells you exactly what is worth changing.

Link to fatigue. A rise in the number of partial responses and a falling completion rate often go hand in hand with survey fatigue. Analyzing partial data helps confirm the hypothesis and shorten the survey or add logic jumps.

How it works in SurveyNinja

In SurveyNinja, incomplete responses are saved and available in the reports section. For more on how to review and use partial data, see the help article "Incomplete responses". From this data you can see how many respondents stopped at each question and export partial responses to analyze the funnel and improve the survey. Totals for completed and incomplete responses are easy to review in the survey summary before exporting or doing a detailed breakdown.

What to look at in partial responses

Where they stopped. The distribution of the "last answered question" across partial sessions shows where drop-off is highest. Questions with a peak in drop-off are candidates for simplification, relocation, or removal.

How many questions they got through. If most partial responses "break off" at question 3–4 out of 20, the problem may be at the start of the survey (a long intro, a complex first block). If drop-off is closer to the end, accumulated fatigue is likely.

Time to drop-off. If the platform records time, you can estimate how long respondents "held on" before leaving. This adds to the drop-off picture.

Re-invitation. In some scenarios, respondents with partial responses can be invited to take the survey again (for example, by email). This is not always appropriate, but for important surveys it can sometimes raise the response rate and the completion rate.

Link to other metrics

Partial responses are the basis for calculating the completion rate and the abandonment rate: "starters" = completers + those who have a saved partial response. Without saving partial responses, "starters" can be underestimated (those who left are not counted), and the completion rate will be inflated. In SurveyNinja reports, the number of incomplete responses is shown next to completed ones; a summary view of the survey is available in the "Reports and responses" section.

Exporting partial responses

Partial responses in SurveyNinja can be reviewed in the interface and, if needed, exported together with completed ones. Export in tabular form (CSV, XLSX) lets you build a "last answered question" funnel in Excel or BI tools, calculate completion and dropout rates by your own rules, and pass the data into a report. For details on exporting, see the "Reports and responses" section. Before exporting, make sure that saving incomplete responses is enabled in the settings, otherwise the export will contain only submitted data.

Common mistakes

Not using partial data. Ignoring partial responses and looking only at completed ones means losing information about drop-off and "bottlenecks".

Confusing them with completed ones. In analysis and reporting you must clearly separate them: completed responses go into content-based calculations; partial responses go into funnel, completion, and dropout analysis.

Not stating it in the methodology. In a report it is useful to briefly note: "Incomplete (partial) responses were taken into account to calculate the completion rate and analyze drop-off by question".

Counting only completed responses in content analysis. For calculations based on answers to questions (percentages, averages, cross-tabulations) use only submitted responses. Partial responses are only for the funnel and drop-off.

Forgetting about methodology. In a report or when handing data to a client, briefly describe how partial responses were taken into account and how completion and dropout metrics were calculated. This increases transparency and trust in the numbers.

Practical recommendations

Enable saving of partial responses. If SurveyNinja settings let you enable saving incomplete responses — use it for funnel analysis.

Review drop-off by question regularly. Use partial data to identify the questions with the highest drop-off and, one by one, simplify or shorten the survey.

State it in the report. In the methodology you can write: "Incomplete responses were saved and used to calculate the completion rate and analyze drop-off points".

Usage example

A survey of 15 questions: 400 people started, 260 completed, 140 — partial responses. The report on incomplete responses shows that 80 of the 140 stopped at question 8 (a complex table). The fix: simplify question 8 or split it into two simple ones. After the change the completion rate rose — partial data helped find the cause of drop-off.

In long surveys, partial data is especially valuable: without it, it is hard to tell whether drop-off is linked to fatigue toward the end or to a specific "heavy" question in the middle. Regularly reviewing incomplete responses in SurveyNinja helps keep the survey in shape and reduce the dropout rate. Before exporting or reporting, it makes sense to reconcile the number of completed and incomplete responses so that the methodology for calculating metrics is transparent.

When reporting to a client or in research methodology, briefly state: whether partial responses were taken into account, how the number of "starters" was determined, and how the completion rate and the dropout rate were calculated. This increases trust in the numbers and shows that the survey funnel was analyzed deliberately.

Bottom line: saving partial responses gives the full survey funnel and accurate completion and dropout metrics; in SurveyNinja this option is configurable and shown in reports.

Partial response — answers entered in a survey but not submitted (the session was not completed). In SurveyNinja, such responses can be saved and used to analyze drop-off by question, calculate completion and dropout rates, and improve the survey. For details, see the help article "Incomplete responses".