It measures how people felt last quarter, which is a different question from whether they can deliver next quarter's plan.
Why it matters when the plan changes
Sentiment data is worth having and is routinely asked to do work it cannot do. A score against a benchmark tells a leader that a unit is unhappy; it does not tell them which decision has no owner, which dependency has no sequence, or which interface the plan runs through. Those are questions about real ownership and dependency, not experience, and research on strategy execution found only nine percent of managers say they can rely on colleagues in other functions and units all the time, the kind of cross-unit reliance a sentiment score never measures.
The tension is cadence and consequence. An annual instrument answers a question that moves monthly, and the action it produces is usually a local initiative rather than a structural change. Self-report is also shaped by social-desirability bias, the tendency to answer favourably, so even the sentiment recorded is not a clean read. Organisations often end up with a well-measured mood and an unexamined plan, connected in no report.
In practice
A division scores well on its annual survey and misses its plan by a quarter. Both are true. People were reasonably content and the plan depended on a handoff between two functions that no one owned. Nothing in the survey instrument was designed to find that, and the good score delayed the search.
Evidence
Reliance across functions is far weaker than within a line, and it is the cross-unit commitments that decide delivery.
Donald Sull, Rebecca Homkes and Charles Sull, Why Strategy Execution Unravels and What to Do About It, Harvard Business Review (2015)Responses to self-report instruments are shaped by the tendency to answer in ways viewed favourably.
Social-desirability bias, Wikipedia (2026)
What it cannot tell you
An engagement survey cannot show whether the decisions a plan depends on have an owner, or whether the dependencies between teams are sequenced. It measures sentiment at a point in time, gathered annually, so it is silent on the structural conditions, ownership, dependency, resourcing, that determine whether a plan is delivered.
Questions
Reported sentiment about work, management and the organisation, aggregated by unit and compared with a benchmark or a previous wave. It is a real measurement of how people say they feel, subject to the usual distortions of self-report in a setting where the employer is asking.
Because the instrument asks about experience and delivery depends on structure. Clear ownership, sequenced dependencies and resourced priorities determine whether a plan lands, and a unit can be content while all three are missing. The two measures are answering different questions.
Yes, for what it is. It surfaces problems people will report, tracks change over time and gives employees a channel. The error is treating it as an organisational diagnostic for execution, which is the use it is most often put to because it is the data that exists.
The most documented is social-desirability bias, described by Wikipedia in 2026 as the tendency to answer questions in a way that will be viewed favourably by others, particularly acute when the employer is the one asking. Aggregation, timing and external benchmarks add further distortion.
Evidence about the plan itself: who holds the decisions it depends on, which interfaces carry the critical path, whether priorities are resourced. Harvard Business Review's 2015 research on strategy execution found only nine percent of managers trust colleagues in other functions completely, which is where delivery usually breaks.