Can I close a survey early and release all outstanding results?
As a requester, you can close your survey early — before it has used up all its response slots. Doing so immediately publishes a final report based on the data collected so far.
This action has important privacy limitations, and it is permanent.
Because of our privacy guarantees, closing a survey early can permanently exclude some data from the final report.
Once a survey is closed early:
- it cannot be reopened,
- any unused slots are forfeited, and
- any data that cannot be safely published is lost from the report forever.
What happens when I close a survey early?
As your survey runs, as a requester, you may see results arrive incrementally in privacy-safe groups in your dashboard. These partial releases are deliberate and only occur when it is safe to do so.
(Read about the “safe batches of three” rules)
When you close a survey early, we evaluate whether the responses that have not yet been released can be safely published, given what you have already seen. Our goal of course, is to release everything ... if we can.
Depending on the situation:
-
All collected responses may be included.
If the remaining responses can be published together as a sufficiently large group, the final report includes everything collected so far. -
Some responses may be held back.
If closing early would cause only one or two new responses to appear on top of data you’ve already seen, those responses are withheld to prevent easy identification. -
Early close may not be available yet.
A minimum of 4 responses is required to produce a report at all. If too few people have responded, early close isn’t an option yet.
All participating respondents still get access to the report
Salary Confidential’s reciprocity promise holds even when a survey is closed early.
Respondents participate as peers and have no control over whether a requester decides to close a survey early.
All respondent result keys remain valid. Any participant who contributed data will have access to the final report as promised, whether or not their individual response appears in it.
Small surveys (4–6 slots purchased): nothing is held back
If your survey ends up with 4, 5, or 6 total responses, all available responses will always be included in the final report — even if you close the survey early.
Why this is safe:
- In this range, the requester has not seen any full rows of data, but instead was given what we call the 'tiny peek' of how the survey had been trending (median and averages of the first 3 responses only)
When is data held back? Can it be recovered later?
In rare cases, up to two responses you've already collected may be held back from the final report when a survey is closed early.
If you'd like to dig into the complete logic, see
How the early close release function works (math edition).
This is because the requester (not respondents) have already seen partial results. Revealing a very small additional batch could make those new responses easy to infer and would break our privacy guarantees.
Held-back responses cannot be recovered later. Closing early is final: you cannot reopen the survey to collect more responses or “complete the batch.” Any unused slots are forfeited.
--> When you click "Close early" in your dashboard, we will show you a confirmation screen which will calculate at the current state of response collection of your survey, if any responses would be held back from your report. You will always have the option to not continue with closing the report early, and instead proceed with collecting the full planned number of responses so you can get a final report with all results.
What does the final report look like?
An early-closed survey report looks the same as a survey that ran to completion.
All calculations — medians, averages, distributions, and ratios — are based only on the responses that were safely published. Held-back data is not included in any metric.
If the survey is part of a poll group, it behaves like any other completed survey and rolls up into poll-level results once at least two surveys in the poll have closed.
Why is Salary Confidential strict about this?
Respondents never see partial data from your survey — only the final report.
Requesters, however, do see results as they are released.
That means our privacy model must account for what you have already learned, not just what appears in the final report. Early close applies the same privacy rules retroactively, rather than relaxing them.
“But this is my data — aren’t you trapping it?”
Responses are processed inside a privacy-preserving system and remain escrowed until release conditions are met. Until then, they are not available in a form that a requester can access, export, or control.
As explained in our Terms of Service, Salary Confidential sells access to a platform and its privacy-preserving infrastructure, not a guaranteed quantity of extractable data.
Some data may therefore never be released. This is not a failure of the system — it is a safeguard that respondents rely on when choosing to participate. And in fact, it is the product promise.
Trust over flexibility (and revenue)
If strict anonymity guarantees are not essential for your use case, there are many excellent tools that give full control to the data owner — from simple forms to advanced polling platforms. These are good tools, but we offer something different: a brokered environment of trust, where respondents participate precisely because requesters agree to give up some control over timing, release, and enforcement in exchange for stronger privacy guarantees.
We could encourage requesters to buy more slots than they need, offer bulk discounts, and say:
“Don’t worry — you can always close early once you have enough data.”
But we don't do this
Instead, surveys start at a size we believe provides a reasonable anonymity baseline. Additional slots can be added only while the survey remains open, encouraging intentional expansion rather than defensive overbuying.
From a business perspective, this limits upsell.
From a trust perspective, it is essential.
<Somewhere, a CMO cries>
Read more about the privacy math (if you’d like)
These articles provide context into the technical approach we take to make the data more resistant to identification and de-anonymization: