Temperature Check

When to use

Top of a retro, planning meeting, or workshop where you want a numeric mood read. Useful when you want to track a trend over a few sessions. Skip it if you've already run two temperature checks on the same scale with the same prompt — by run three, the team converges on a safe middle number and you've measured nothing. Use ESVP instead when you want a categorical decision input today; use one-word check-in when you want a quick verbal read.

How it runs

1. Pick the scale before the meeting

   1-5 for teams under eight — forces a position. 1-10 for groups of twenty plus, where the finer scale gives the histogram shape. Don't drift between scales meeting to meeting; the comparison breaks.

2. Ask one specific question

   'How are you?' is too vague. 'How confident are you we'll ship by Friday?' is a different format — you've turned a check-in into a planning poker. Pick a single specific dimension (mood, energy, confidence) and stick to it for the round.

3. Vote anonymously, simultaneously

   Poll bot, hidden Miro voting, sticky notes face-down. Public temperature checks anchor on whoever speaks first; you'll get a clustered histogram that reads as agreement and isn't.

4. Show the histogram, name the outliers

   A 4 in a room of 7s and 8s is the entire point of running this. Don't ask the 4 to explain — ask what would help. Naming the outlier directly is the only useful response; pages that stop at 'discuss findings' missed the format's job.

5. Rotate the question or retire the format

   The third use of the same prompt produces convergence on the safe number. Change the question (mood → confidence → energy) or stop running it for a few weeks. The format only earns its place if the data still moves.

Why it works

A number forces commitment that words don't. Asked 'how are you?' people answer 'fine'; asked to pick between 1 and 5 they pick 3 if they mean something other than fine. The histogram surfaces distribution — a team averaging 6 might be five 6s or one 9 and four 4s, and the second team needs a different meeting. Anonymity is the format's signal-to-noise ratio; without it you measure conformity, not mood.

Variations

• 1-5 with named labels (1 = struggling, 3 = okay, 5 = great). Useful for teams new to numeric scales.
• Two-axis temperature: 1-10 on workload, 1-10 on focus. Two histograms, slightly more diagnostic.
• Async temperature in a Slack poll the morning of the meeting. Lower live cost, slightly weaker engagement.
• Trend-tracked: log the team average each retro. Three sprints of declining average is a real signal.

Facilitator notes

Vote yourself, last, and pick a number that's not the obvious median. If you always vote 7, the team learns 7 is the safe answer. If your honest answer is 4, vote 4 and say so — the format dies if the facilitator games it. Don't ask outliers to justify their score; ask what would move it up by one.

Pitfalls

• Public votes. The first number anchors the rest; you measured conformity.
• Drifting between 1-5 and 1-10. The trend across meetings becomes uninterpretable.
• Same prompt for the third meeting. Everyone votes 7; nobody's lying, the format's just dead.
• Stopping at 'discuss findings.' Outliers are the data — name them or you ran a survey.

Remote tips

Anonymous poll bots (Polly, Slack polls, Mentimeter) work better than reaction emojis — emojis read public, even in threads. Show the histogram on shared screen; numbers in chat without a visual lose half the signal.

Example outputs

• Histogram on a 1-10 scale: 8, 8, 7, 7, 8, 9, 4. The 4 is the meeting now — name it directly.
• Three weeks of averages: 7.8, 7.2, 7.5. Flat. Either the team is genuinely steady, or you've trained them to vote 7 and the format is dead.
• 1-5 confidence on shipping by Friday: 4, 4, 3, 2, 4. Two-thirds confident, one person flagging real risk. Stop the standup, hear the 2.

FAQ

Temperature check or ESVP?

Temperature check is a 1-10 numeric scale and works for trend tracking across meetings. ESVP is a four-label categorical (Explorer/Shopper/Vacationer/Prisoner) and works for a single-meeting decision input. Use temperature check when you want a number you can graph; use ESVP when you want a category that tells you whether to pivot the agenda.

1-5 or 1-10?

1-5 for teams under eight — the forced choice produces sharper data on small samples. 1-10 for groups of twenty plus, where the finer scale lets the histogram show shape. Pick one and stick to it; mixing scales meeting to meeting kills the trend.

What do I do with a 3 or a 4?

Name it. Don't ask the person to justify; ask what would move their score up by one. The actionable version of a low score is the 'what would help' answer, not the score itself. A 4 followed by silence is the format failing.

Related activities

Recommended use cases