Engagement cycles

An engagement survey for the whole organization, bought once per cycle.

A cycle asks everyone in the organization the same questions, once, and reports what they said as group numbers with error bars — for the organization, for each unit, and for each manager. Groups too small to report are left out rather than quietly included, and so is any group that would give one away.

  • ~10 minutes per person
  • 40 questions + 2 open
  • Reporting floor: 5 people
  • From $999 per cycle
  • Bought once, not monthly

Setting up costs nothing and needs no card. Everyone answers from a personal link; nobody but your admins signs in.

What people answer

The team you work in

The same 34 core statements the team survey uses, asked about the team somebody actually works in day to day. That is deliberate: it means a unit’s numbers mean the same thing whether they came from a cycle or from a team survey, and it lets every unit’s climate figure carry the same agreement test the team survey applies — the scorecard says when a group’s members do not describe the same team. A questionnaire that asked about “the organization” instead would produce a number nobody can act on.

Your own work

6 questions about effort, meaning, absorption, whether somebody expects to still be here in a year, and how likely they are to recommend the place. Then 2 optional open questions — one about what would make it better, one about what is working and should not change.

The items are ours, written to published definitions rather than licensed from anyone, and they are published: the instrument page shows each block’s first dimension to anyone, and one emailed link opens every question, the scoring rules and the thresholds before you buy.

And your own questions, kept separate

You can add up to 10 questions with a fixed answer — pick one, pick several, a rating on a scale you label, or options put in order — and up to 3 open ones, at no extra cost. Ask them once, or mark them to return every cycle and watch them move. They are asked after everything above, never before: a question about a new policy asked early colors every answer that follows it, and the blocks above are the half of the survey that has to stay comparable.

They come back under their own heading, never mixed into the scorecard and never described as validated — they mean what you wrote them to mean. Once a cycle carrying a question has opened, its wording is fixed: editing it makes a new question that starts its own trend, because two wordings are not two readings of one thing.

What comes back

One scale of an engagement report, by unitFive report rows on a one-to-five scale. The organization and two units show a mean with a 95 percent interval and the number of respondents; one unit is withheld for being under the floor of five, and a second is withheld because publishing it would reveal the first by subtraction. Illustrative values.Direction and role clarity · by unit12345Organizationn = 2123.9Unit An = 414.1Unit Bn = 373.6wider: members disagreedUnit Cn = 4withheld — below the floor of 5Unit Dn = 9withheld — would reveal Unit C by subtractionIllustrative. Every published mean carries its n and interval; a group under the floor is withheld, and so is a sibling that would let you subtract it back out.

Ask everyone once. Get numbers you can defend. Every mean comes with the number of people it came from and a 95% interval. A wide interval means the people in that group disagreed with each other, or that there were few of them — and either way it is a reason not to act on a difference of 0.3 between two teams of six. Most reporting shows you the average and stops.

A written summary of the numbers is produced for you by a language model that is given the published group figures and the themes, and never anybody’s individual answers. The numbers themselves are calculated, not generated.

Numbers with error bars

Groups too small to report are left out. So, sometimes, is a second group beside them: when a division publishes a number and all but one of its teams do too, the missing one is recoverable by subtraction, and holding back only the small team would be a promise we did not keep. That is the part that is easy to get wrong, and it is why a team you expected to see is occasionally absent.

The threshold is five by default. You can raise it. You cannot lower it below five, and we will not tell you it is a legal requirement, because as far as we can establish there isn’t one — the question regulators actually ask is whether somebody can be identified, which is not the same as counting to five. Nothing here is legal advice. How reporting thresholds work, and the subtraction problem.

What people write stays theirs

The open questions are optional, and what somebody writes is never shown to anyone at their organization. Not to an administrator, not in an export, not in a report. What is reported is themes — and only themes that at least three different people raised, written in our words rather than lifted from anyone’s.

Anonymity is an effort, never a guarantee. Answers are tied to a person so somebody can finish later and nobody can answer twice — which means keeping them private is something we do on purpose and can explain, rather than something that is impossible by accident. We tell participants that too, in those words, before they start.

A run is a team. A cycle is an organization.

DimensionTeam survey (a run)Engagement cycle
Who answersOne team of 3–12, about each other and the teamEveryone in the organization, about their own team and their own work
What they rateNamed teammates' observable behavior + team climateTeam climate + their own engagement; nobody rates a person
What comes backTeam pack, a private note per person, a facilitator view, a 60-minute debrief agendaGroup numbers with intervals for the organization, each unit, each manager; themes from open text
Smallest reportable group3 raters for anything, 5 for question-level detail5 by default, raisable; subtraction-safe
Price unitPer run, per teamPer cycle, for the whole organization, by headcount band
DebriefIncluded, alwaysNot produced — a team that wants one runs a team survey

The two share an instrument on purpose, so a unit’s climate numbers mean the same thing wherever they came from. The full comparison, including 360s.

What it costs

Bands

EmployeesFirst cycle in a yearSecond cycle, same year (60%)Top-up block
Up to 100$999$59910 seats, $100
101–250$1,799$1,07910 seats, $75
251–500$2,999$1,79910 seats, $60
501–1,000$4,999$2,99910 seats, $50
1,001–2,500$9,499$5,69950 seats, $190

Practitioners — on the Practitioner plan or certified — pay 25% off list; the two discounts do not stack.

How top-up blocks work

Between bands you can add blocks of ten or fifty seats at the rate the band you are already in charges, so a boundary is a step of 1.09× rather than the 1.78× it would be without them. At 100 people a cycle works out around $9.99 a head, and at 101 it is around $10.88.

Blocks come in ten or fifty seats rather than ones on purpose. A per-person number would be a headcount we track and a bill that moves with it, which is exactly what this pricing exists to avoid. So one extra person buys a whole block — and you keep the rest of it. Above a thousand people a block is fifty seats rather than ten, because at that size ten-seat blocks stop being a purchase and start being an inventory. Past a point the next band up is simply cheaper, and you get that instead without having to work out which is which.

Above 2,500 people

An Enterprise agreement, with published prices from $25,000 a year. What changes at that size is not the arithmetic but the work — the questions in more than one language, a roster fed from your HR system rather than a spreadsheet, works councils consulted before anyone is asked anything, and a decision about how many units you actually want reported. None of that is something a checkout can do, which is why that size is quoted.

Who sees what

Workspace owners and admins

They run the cycle, because the roster is everyone. They see every reportable group’s numbers, the themes, and the written summary. Not the open-text answers — those are shown to nobody at the organization.

A manager

Appears in the report as a group: their unit’s numbers, once the unit clears the floor. Never an individual answer, never a name.

The people who answered

Their own answers, while the cycle is open. Nothing they wrote is ever shown to anyone at the organization.

Questions

How is this different from the team survey?
Scope. A team survey is one team of 3–12 rating each other, ending in a debrief. A cycle asks everyone in the organization the same questions once and reports group numbers by organization, unit and manager — no debrief pack, no individual notes.
What do people actually answer?
The standard survey is two blocks: the same 34 core climate statements the team survey uses, asked about the team someone works in day to day, then 6 questions about their own work and the organization, and 2 optional open questions. About ten minutes.
Can we add our own questions?
Yes, at no extra cost: up to 10 with a fixed answer and 3 open, one-time or every cycle. They are asked last and reported under their own heading, never mixed into the validated scores. Once a cycle has asked one, the wording is fixed.
Can a manager see who said what?
No, and it is not a setting. Individual answers have no read policy for any signed-in account; results appear only as group numbers. Open-text answers are never shown to anyone at the organization — only themes several people raised, in our words.
What happens to a small team?
It is left out of the report rather than shown. Sometimes a second group is left out beside it, because if a division publishes a number and all but one of its teams do too, the missing one can be worked out by subtraction.
Is the minimum group size a legal requirement?
Not that we can find, in any jurisdiction we have looked at. Regulators ask whether somebody can be identified, not whether you counted to 5. Five is our default because it is defensible; you can raise it, not lower it. Not legal advice.
Do you have benchmarks against other companies?
No, and we are not planning to. A benchmark places you against a sample you cannot inspect, drawn from companies that chose to be in it. We show your own numbers with honest error bars so you can compare yourself to yourself next time.
What does it cost to run twice a year?
A second cycle in the same calendar year is 60% of list — at 101–250 employees, $1,799 then $1,079. Practitioners get 25% off list. The two discounts do not stack.
We have 108 people. Do we pay the 101–250 price?
No. You pay the band below plus one block of 10 seats, priced at that band's own rate, so cost per person barely moves across a boundary. Past a point the next band is simply cheaper, and you get that instead.
Can we run it for more than 2,500 people?
Yes, as an Enterprise agreement with published prices from $25,000 a year. What changes is the work — more than one language, a roster from your HR system, works councils consulted — not the arithmetic. See /enterprise.

Set up a cycle

Upload a roster, pick the band, and everyone gets a link. Results by organization, unit and manager, with the small groups withheld — and nothing charged until you open the cycle.