Overview
This guide covers what makes a binary check-in question worth asking. By the end, you'll have a question that generates honest signal and a clear sense of what to do when the scores move.
Start with something you can act on
The test for a good check-in question is not whether it sounds insightful. It's whether a run of 'no' answers tells you what to do next.
'Are you feeling engaged at work?' is hard to act on. A dip in results could mean anything from a difficult sprint to someone edging toward the door. You can't tell, so you can't respond. That's not a signal. It's noise with a percentage attached.
Try this instead: 'Did you feel clear on your priorities this week?' A sustained drop points at something specific. You know what to raise in the next 1:1. You know what to change at the team standup. The question creates a path forward rather than a score you're left staring at.
Start narrow. A question about one specific, operational thing generates a trend line you can read. Broad feelings questions generate sentiment. Sentiment is useful context; operational signal is what drives action.
The anatomy of a good binary question
Three things to check before you publish.
It asks a single thing. 'Were you productive and motivated this week?' is two questions joined together. If one answer is yes and the other is no, the respondent has to average them. The result is meaningless. Split it, or choose the one that matters most to you right now.
It's answerable in two seconds. Relay is built for a 30-second interaction. If your question makes someone pause to evaluate three months of their working relationship, it's doing too much. Questions about culture, values, or alignment tend to abstract their way into slow, ambiguous answers. Questions about specific conditions this week tend to be immediate and honest.
A 'no' is informative. The best check-in questions are ones you could act on if the score drops. 'Did you feel clear on your priorities this week?' drops: have a conversation about direction. 'Did you have uninterrupted time to focus?' drops: look at the meeting schedule. 'Did your work feel meaningful this week?' drops: you need to understand whether that's about recognition, scope, or something structural. In each case, a low score points somewhere.
A few questions worth adapting: 'Did you have what you needed to do your best work this week?' 'Did your work feel meaningful this week?' 'Were you able to focus without unnecessary interruptions?' 'Did you feel clear on what success looks like this week?' None of them are perfect for every team. All of them are answerable in two seconds and actionable when the trend shifts.
What to avoid
Leading questions generate noise. 'Was this a productive week?' tends toward yes regardless of reality. Nobody wants to flag themselves as unproductive in a check-in their manager reads. The anonymity helps, but it doesn't fully override a question framed to expect a positive answer.
Questions that require explanation are usually the wrong question. If you find yourself adding a note in brackets ('Did you feel supported? By your manager, your team, or cross-functional partners'), that parenthetical signals you're trying to ask three things at once. Pick the most important one. If you're tempted to explain, you probably already know which part you're actually worried about. Ask that directly.
Vague emotional territory is genuinely harder to act on. 'Did you feel good this week?' might capture anything from a difficult client call to a bad commute. That's not useless, but it doesn't tell you where to look. Save the broader wellbeing questions for after the habit is established and you have a baseline to read movement against.
When to update the question
The question can be updated at any time. But changing it too often makes the trend data meaningless. If you swap questions every third week, you can't tell whether anything is getting better. You just have a series of disconnected scores.
Run the same question for four to six weeks before reviewing it. By the end of that period, you'll know whether it's generating useful signal: the scores are moving in a direction, the comments point at something, and you're reading the results differently than you were in week one.
At the review, ask one question: has this told me something I could act on? If yes, keep it for another four to six weeks. If every result has landed in the same narrow band and the comments are sparse, the question is either too comfortable or too abstract. Tighten it.
You don't need the perfect question to start. You need one that generates a signal you can act on. Write it, run it for a month, and let the results tell you what's missing.