Retrospectives are where a team decides what to change next sprint, so the order people speak in is not a cosmetic detail — it shapes the outcome. If the same person opens every retro, their read of the sprint becomes the reference point everyone else reacts to. Drawing a random, even, no-repeat order at the start hands that framing role to a different person each time, and guarantees everyone a turn.
Why order matters more in a retro than in stand-up
A stand-up is a quick status sync; a retrospective asks people to be candid about what went wrong and what to try. Three well-documented effects make the first voice powerful in that setting. Anchoring means the first problem raised becomes the number everyone adjusts from. The serial-position effect means what is said first (and last) is remembered and weighted more than what came in the middle. And once a senior or highly confident voice frames the sprint, others tend to fall in line rather than open a competing thread — the pull toward groupthink. A random order does not remove these effects, but it stops any one person owning the powerful first slot every single sprint.
Random, but still fair
Random only helps if the draw is even and does not quietly repeat. WheelMigo draws each pick from your browser's crypto.getRandomValues with rejection sampling, so no name is favoured, and no-repeat rounds mean everyone speaks once before anyone speaks twice. We checked this the boring way: across a test of seven million draws the distribution stayed flat, with no position winning more than chance.
Set it up once, reuse it each sprint
Retros are periodic, not daily, so the setup should survive between them. Add your squad, save it as a roster, and it is there for the next retro and the one after. When someone joined mid-sprint, add them; when someone is on leave or in another meeting, mark them as not present so the wheel skips them without deleting them from the team. Nothing is uploaded — the roster lives in your browser on your device.
Reshuffle for each prompt, not just once
This is where a retro differs from a single stand-up order. A retro usually runs several prompts — what went well, what did not, what to try next — and each one benefits from a fresh order. If you keep the same order across every prompt, the person drawn first still opens the whole session. Spin a new order for each prompt: within a prompt, no-repeat rounds give everyone a turn; between prompts, a reshuffle means a different person opens each topic.
| Retro moment | What to use | Why |
|---|---|---|
| Setting the stage / check-in | One no-repeat round | Everyone speaks once, in an order nobody can predict |
| Each data-gathering prompt | Reshuffle, then a no-repeat round | A different person opens each topic, so no one frames them all |
| Voting or deciding actions | No wheel — discuss openly | Prioritisation needs judgement, not a draw |
For remote and hybrid retros
On a video call, turn on Present mode: it hides the editor and shows just the wheel, large enough to read on a shared screen. Remote retros are exactly where a visible, obviously-fair draw helps, because people cannot see the room and are quicker to assume the facilitator is picking favourites.
What a random order will not do
A fair order is one lever, not the whole retro. It will not create psychological safety on its own, it will not make people candid in a culture that punishes candour, and it does not replace a facilitator who draws people out and keeps the conversation useful. What it does is remove one small, recurring source of unfairness — the same person always going first — so the facilitation you do add lands on a more level start.