You Said, We Did: Closing the Loop on Anonymous Feedback

01The loop most surveys never close

An anonymous survey has a structural problem that a named one doesn't. You can't thank anyone. You can't tell the person who raised the parking issue that it's fixed.

You can't follow up with the team that scored lowest, because you don't know which individuals scored low. The respondent's experience ends at "thanks for your response" and then, usually, silence.

Silence is expensive. The article on why participation rates stall around 60% makes the case that a large slice of non-response is rational: people who don't believe the survey is anonymous, and people who don't believe it changes anything. Encryption addresses the first group.

Nothing in the architecture addresses the second. Only what you do afterwards does.

So the reply has to be public and collective. Not "we heard you, Sam", which you can't say, but "you said this, we did that", to everyone who was asked.

02Why three, not a project plan

The competitors in the pulse category all ship some form of action planning, and the enterprise versions are project tools: owners, due dates, progress bars, reminders. We built something smaller on purpose.

An action plan in InviziPoll is up to three commitments, each with a title, an optional line of detail, and a status: planned, in progress, or done. That's it. No owners, no dates, no notifications.

Three is enough for the thing that actually moves the next response rate, which is a respondent opening the next pulse and seeing that the last one led somewhere. Ten actions with owners is a management artefact. Three with a status is a message to the people who answered, and messages should be short.

03Where it shows up

The admin records the actions on the poll's results tab and publishes them when ready. Nothing is visible until publish. After that, the plan appears wherever a respondent might look:

• On the poll's own link once it closes. Someone who opens a closed poll sees what came of it instead of a dead end.
• On the public results page, above the charts, if the poll has one. It shows there even before numbers have been published, because "we've committed to this, results are still coming" is a legitimate thing to say early.
• At the top of the next wave's response page, if the poll is one wave of a recurring series. A short "last wave, you said, we did" strip sits above the new questions, which is the exact moment the theory says it matters.
• In Slack, if the poll was announced in a channel. One button posts the published actions to that channel, with the results link if public results are on.

Every one of those surfaces carries the same sentence: nothing anyone wrote is included. The actions are the organisation's words about what it will do. No verbatim, no count, no chart travels with them.

04What the loop does to anonymity

Nothing, and that took some care.

Responses in InviziPoll are encrypted in the respondent's browser and the server stores ciphertext it can't read. Action plans are the opposite: plaintext, like a poll title, because they're meant to be broadcast. Encrypting them would break the Slack post and the next-wave strip for no privacy gain, since they contain nothing from any respondent.

There's also no response threshold on publishing actions. Results wait for the poll's minimum group size, never fewer than three, before anything shows, because a chart over two people identifies them. Actions are the admin's own words about their own workspace, so the gate would protect nobody. You can publish "we heard you, we're on it" the day the survey opens.

And they're not tier-gated. Free workspaces and expired trials can author, publish and serve an action plan. Gating the loop-closing half of a product whose pitch is "answers don't vanish into a void" would be perverse.

05What this looks like across a series

The place it pays off is a recurring pulse. Each wave's actions are published to that wave's respondents, and they're the strip the next wave's respondents see before they answer. On the series trend page, each wave shows the actions that were published for it beside the movement in the numbers, so a colleague reading the trend can see what was promised alongside what moved.

That's the honest form of "closing the loop". Not a claim that action planning improves engagement, which is the kind of sentence that appears in vendor decks unsupported. A record, wave by wave, of what was said and what was done, next to what happened to the numbers.

06Doing it well

A few things we've seen make the difference between a plan people notice and one they don't.

Publish something before the results are final. "Results are still coming; here's what we're already doing about last time" is worth more than a polished plan six weeks later.

Move the status. Editing after publish keeps the plan published, so you can move an item from planned to done without republishing. A plan that never changes status reads as decoration.

Be specific and small. "Improve communication" is not an action. "Weekly written update from the leadership team, starting Monday" is. If the AI summary drafted a suggestion you agree with, there's a button that turns it into a draft row, and the draft is yours to sharpen.

Post it where people already are. The Slack update exists because a plan nobody sees closes no loop.

The how-to is in You said, we did: publishing actions after a poll, and the participation argument is in Why your engagement survey participation rate is stuck.

For related guidance, see confidential vs anonymous surveys and how to run an anonymous employee survey.

FAQQuestions people ask