Skip to content

Pulse survey

A pulse survey is a short questionnaire run frequently, typically monthly or quarterly, to track sentiment between longer instruments. The short cycle is the point: it catches movement a yearly instrument would miss. It shares the limits of any self-report about feeling, and adds fatigue when nothing visibly follows.

Frequency turns a snapshot into a trend, and it also turns a question into a routine people stop answering carefully.

Why it matters when the plan changes

Movement is more informative than level. A unit scoring moderately and falling is a different situation from one scoring moderately and rising, and only a short cycle can tell them apart. That is the genuine contribution, and it is why pulses replaced annual instruments as the default in many organisations. Signal only earns that status when it sits alongside role data, strategy documents and other evidence rather than standing alone as the whole picture.

The tension is the response loop. Asking frequently creates an expectation that something happens, and when nothing visibly does, response rates fall and the answers that remain come from the least representative people, coloured further by the tendency to answer in ways that look favourable to others. An instrument with no consequence attached degrades faster the more often it is run, so a short cycle raises the cost of collecting and filing rather than acting.

In practice

A company runs a monthly pulse for two years. Response rates fall from eighty per cent to thirty-five. The remaining respondents are disproportionately those with strong views in either direction. The trend line continues to be presented to the board as though it described the organisation, and by then it describes a self-selected third of it.

Evidence

What it cannot tell you

A pulse measures self-reported sentiment at a moment in time; it cannot on its own distinguish a short-term dip from a structural problem, or a genuine shift from social-desirability bias in what people are willing to say. It says nothing about the causes behind a score, only that a score has moved.

Questions

When response rates start falling and nothing visible has followed previous rounds. Social-desirability bias, described on Wikipedia in 2026, means the respondents who remain are more likely to answer in ways that look favourable, so a shrinking pool is a signal in itself, not just a data-quality problem.

A small number of items that are stable across waves so the trend means something, plus at most one or two rotating items. Changing the question set between waves destroys the comparability that was the reason for running a short cycle in the first place.

Over whatever threshold protects the individuals in it, commonly five or more and higher for sensitive items. Small teams are exactly where managers most want the data and exactly where reporting it identifies people, especially when the group changes between waves.

It can flag movement worth investigating, but it does not identify structural causes. Donald Sull, Rebecca Homkes and Charles Sull found in 2015 that only nine per cent of managers say they can rely on colleagues in other functions and units all the time, a gap no self-report about mood can surface.

Asking without acting, asking the same thing without explaining why, and allowing response rates to fall without treating that as a finding. A falling response rate is itself information about the organisation, and it is usually reported as a data-quality footnote instead.