Auto-save (automatic saving)

A respondent started a long survey. 15 minutes of work, 20 answers. They accidentally closed the tab, reloaded the browser, or just got distracted by a phone call.

Without auto-save, it's all gone — reopening the survey means starting from scratch. Most likely the person simply won't come back: there's no motivation to repeat 15 minutes of work. Auto-save solves this: progress is saved as the form is filled in, and the respondent can return to exactly where they left off.

Definition

Auto-save (automatic saving) is a feature that automatically saves a respondent's answers while they take a survey, without requiring any manual action or submission of the whole questionnaire. It lets them continue later from the point where they stopped. It reduces answer loss caused by technical failures, device switching, and distractions. It is directly connected to working with partial responses.

How auto-save works

The mechanics depend on the implementation, but the general principle is: with each answer (or group of answers), the data is sent to the server and saved against the respondent's session. When the user returns, the system recognizes them — by cookie, unique link, or login — and restores the state of the survey.

Typical save triggers:

• After each question. Maximum data safety, but more requests to the server
• When moving to the next screen. A balance between safety and load
• Periodically on a timer. For example, every 30 seconds — suitable for long open-ended answers
• When the tab is closed (beforeunload). The last chance to save data before the person leaves

Why auto-save matters

Reducing losses in long surveys. For questionnaires longer than 5-7 minutes, auto-save is critical. Without it, every interrupted session = a lost respondent. With it, the person comes back and finishes filling it in.

Working across multiple devices. Started on the phone on the subway, continued from a laptop at home — for this you need not only auto-save but also a session identification mechanism. It works via a personal link, a login, or saving to a cloud cookie.

Protection against technical failures. A network glitch, a frozen browser, a dead phone — the data that managed to save isn't lost. The respondent returns not to an empty questionnaire but to a partially completed one.

Respondent comfort. The feeling "I can put this off for an hour" reduces the psychological pressure of a long survey. People are more willing to start filling it in, knowing they don't have to do it all in one go.

Auto-save and partial responses

Auto-save creates a partial response — an unfinished answer that already contains part of the data. This raises an important analytics question: how should such answers be counted?

Three approaches:

• Count only completed responses. Partial responses don't enter the analysis. A simple approach, but it loses information — some partials may contain the entire important part of the questionnaire.
• Count partials with a caveat. Answers to the first N questions are counted, and the absence of data for later ones is flagged. This increases data volume but requires careful handling of gaps during analysis.
• Filter by % completion. Answers where more than 70-80% of the questionnaire is filled in are included; less than that — excluded. A compromise between data volume and quality.

The approach is chosen to fit the task. For NPS, any partial with a score is already valuable. For validating a scale with a Cronbach's alpha calculation, you need full completions.

How to bring a respondent back to a saved session

Via cookie. The simplest way: on the first visit a session marker is written, and on the next one it's read, and the survey opens at the saved point. It works as long as the cookie isn't cleared, on the same device and browser.

Via a unique link. In an email campaign or a personal invitation, the link contains a session token. The respondent can close the window and return via the same link from any device — the survey continues from where they left off.

Via login to the system. For corporate surveys where the respondent is authenticated through SSO, the session is tied to the login. Convenient for long HR surveys that people may fill in over several days.

Via email for reminders. If the respondent left an email at the start of the survey, the system can send them a reminder: "You started a survey and didn't finish — here's a link to continue." This significantly improves the final completion rate.

Example: the effect of auto-save on a long survey

A research company runs a 35-question survey about consumer behavior — the average completion time is 18 minutes. They compared two versions:

• Without auto-save: of 1,500 people who started the survey, 420 (28%) dropped out in the middle. All those sessions are completely lost data.
• With auto-save and an email reminder: of 1,500 who started, 380 (25%) dropped out on the first session. But of those, 165 came back via the reminder and finished the survey (43% return rate). Total irrecoverable losses — 215 respondents (14%).

With auto-save and a reminder they recovered 165 answers out of a potential 420 lost — that's 40% more data for analysis. For tools where the cost of acquiring each respondent is high (panels, B2B surveys), this is a direct saving.

Common mistakes

Not telling the respondent that auto-save is working. If a person doesn't know their answers are being saved, they may get extra nervous about possible loss of work or stop filling it in, fearing everything is already gone. An explicit message ("Your answers are saved automatically — you can come back later") reduces anxiety and improves completion rate.

No mechanism to return to the session. Auto-save works, but the respondent doesn't know how to come back. Without a unique link or a reminder, the data is saved into the void: no one will come in to continue.

Storing data for too long after the session starts. If a partial response comes back from someone who started the survey six months ago, they have most likely already taken the survey in another wave, or the data is out of date. A reasonable storage period for unfinished sessions is 30-60 days, after which the session is reset automatically.

Ignoring partial responses during analysis. The opposite extreme: auto-save works, respondents return, but only 100% completions make it into the analytics. In this case you lose part of the data that respondents actually managed to provide.

Auto-save in SurveyNinja

In SurveyNinja, automatic saving of answers is built into the standard settings — answers are saved as the questionnaire is filled in, and the respondent can close the tab and return to the survey via the same link (with the cookie preserved) or through a repeat personal link in the case of an email campaign. Auto-save runs in the background while the respondent works through the survey.

In reports, both fully completed answers and partial responses are available — unfinished, but with saved intermediate data. This makes it possible to use the partial-responses-included approach in analysis: with a low completion rate on long questionnaires, partial responses can substantially increase the volume of usable data.

Auto-save is a standard feature of modern survey platforms that turns "lost" sessions into resumable ones. Without it, a long survey loses 20-40% of respondents to interrupted sessions. With it, a significant share come back and finish — especially with smart use of email reminders. The key elements: reliable data saving, a mechanism to return to the session, and an explicit message to the respondent that progress won't be lost.

Frequently asked questions

How long before a session should "expire"?

For most surveys — 30-60 days after the start. Long enough for a person to come back if they wanted to, but it keeps the data from going stale. For time-sensitive research (seasonal surveys, event evaluation) — 7-14 days. After expiry, the session is reset or marked as finally unfinished.

Should you show the respondent progress and saving?

Yes, both elements affect completion rate. A progress bar ("You've completed 40%") gives a sense of the path remaining. The message "Your answers are being saved" reduces anxiety. Together they increase the likelihood of completion — both on the first session and on the return.

Can answers to anonymous surveys be auto-saved?

Yes — technically auto-save doesn't require identifying the person, only a session identifier (cookie or token in the URL). The data can be stored on the server without being tied to an email or name. This makes it possible to combine anonymity with the ability to come back and finish.

How does auto-save work when switching devices?

By cookie — it doesn't work (the cookie is tied to the browser). By a unique link (token in the URL) or by login — it works: open the same link on another device, and the survey continues from the saved point. This approach is used for a seamless cross-device experience.

What happens if a respondent accidentally submits the survey before finishing?

It depends on the settings: in the standard configuration, after the final submission the survey is considered complete and re-taking is blocked. If this is a mistake, the respondent is either left with no way to fix it or an administrator has to step in manually. For important surveys, it makes sense to add a confirmation on the final screen: "Are you sure you want to submit your answers?"