How do we act on engagement survey results and actually move the score?

Acting on the results is the entire value of an engagement survey, because the questionnaire measures a condition and changes nothing by itself. A systematic review of 53 studies on survey follow-up, covering research from 1952 to 2021, found that organizations frequently treat the survey as finished once the data are collected, which is exactly when any return on it is decided.[1] The work that follows is a small, repeatable sequence: share the results fast and without spin, choose one or two themes you can genuinely influence, give the action to managers at the team level, bring employees into designing the fix, report back on what changed and what did not, then re-measure honestly.

Groups whose results were shared and acted on rated their jobs more favorably in later years, while groups that heard nothing, or heard something and then saw no follow-through, grew more dissatisfied and less willing to take the next survey seriously.[1] Gallup, which sells one of the most widely used instruments, states in its own guidance that asking for feedback and doing nothing tends to lower engagement and raise turnover, and by its own global count engagement sat at 20% in 2025, its lowest since 2020.[2][9]

This answer is the playbook for the action itself; whether surveys are worth running at all is covered in a companion answer on whether engagement surveys work. Here the survey has closed and the results are in front of you.

Share the results quickly and honestly

Get the results in front of employees within a couple of weeks of the survey closing, including the parts that reflect badly on the organization. Gallup advises communicating results shortly after the survey closes, while the feedback is still fresh, because a long silence followed by a polished summary is what employees tend to read as a sign that nothing will happen.[2] People already know where the low scores are, so a report that hides them mainly signals those scores are not open for discussion.

Focus on one or two themes you can actually influence

Pick a very small number of themes and let the rest wait, because attention rather than data is the scarce resource. Culture Amp recommends sticking to one or two focus areas so a team can make meaningful change.[3] A theme also has to sit inside your control: rating pay fairness while the pay bands are fixed for two years produces a number you will publish and cannot move, and CIPD warns that broad composite scores in particular are often very hard to act on.[5]

Hand the action to the team and its manager

Push the work to the level where the rated conditions are actually set, the individual team and its direct manager. The follow-up literature describes the effective pattern as a bottom-up process run in work families, meaning a manager working through their own team's results with that team rather than a head-office committee reviewing data in aggregate.[1] Give each manager their team's results, a simple conversation template, and a clear expectation that the discussion will happen. Gallup calls the manager level the biggest single opportunity to change engagement.[2]

Involve employees in choosing the fix

Bring the people who raised an issue into designing its solution, because ownership and the quality of the fix both improve when it is built with the team. Viva Glint's framework states that involving employees as collaborators in choosing focus areas strengthens ownership and builds trust through transparency,[4] and Culture Amp's guidance likewise has the team brainstorm options before anything is committed.[3]

Close the loop with a visible 'you said, we did'

Tell people what you heard, what you are doing about it, and what you decided against and why. This is the step tied most directly to trust in the research: the follow-up review found that groups who reported their results had been acted on became consistently more favorable raters across items and years, while a recurring source of dissatisfaction was managers who never followed through.[1] Viva People Science notes that employees do not expect every item addressed, as long as they helped set the priorities and can see progress.[4]

Assign owners, dates, and a normal rhythm

Attach a named owner and a date to each committed change, and run it through meetings that already exist rather than a separate initiative. The barriers practitioners cite most for failed follow-up are execution, senior attention, and resources, so a plan with no owner and no deadline is the default way action dies.[1] Track a few leading indicators between surveys, such as attendance at team meetings or movement on a stalled process, so you are not blind until the next measurement.

Re-measure and attribute honestly

Measure again on a sensible cadence, which Gallup puts at a deeper survey roughly twice a year with lighter pulses between, then attribute any change carefully, since a rise can also be unrelated business improvement or plain statistical movement.[2][6]

Days 1 to 10: read the data and share it

Analyze the results and get an honest summary back to employees inside the first two weeks, including the weak spots. Senior leaders and HR own this window: confirm the headline findings and name the one or two themes the organization will prioritize, while the survey is still fresh, which Gallup identifies as the moment a response lands best.[2]

Days 10 to 30: managers meet their teams

Give every manager their own team's results and have them hold a team conversation, treated as expected rather than optional.[1] Managers acknowledge what the data show, invite the team to prioritize among the local issues, and leave with a small number of agreed next steps the team helped choose.[4]

Days 30 to 60: commit and publish the plan

Turn the conversations into a short, owned action plan and make it visible: each change carries a name and a date, declined requests are explained rather than ignored, and the plan goes back to employees as a plain account of what was heard and what will happen.[1] Pick a couple of leading indicators so progress shows before the next formal survey. By day 60 the sequence has moved from a dataset to committed, visible change, the only stage the research associates with a later rise in the score.[1]

Most of what an engagement survey measures is set inside the team, so the person who can change it is usually the direct manager rather than head office. The follow-up literature describes the effective intervention as a bottom-up process carried out in work families, where a manager discusses the team's own results with the team and agrees on local action, and treats managers as the change agents the process depends on.[1] A company-wide initiative can shift things that are genuinely organizational, such as a broken tool or an unfair policy, but it cannot repair how one manager runs one team.

Managers trained to run feedback sessions produced action programs that were more feasible and timely, and were more likely to improve employee attitudes, than managers handed results with no support.[1] CIPD places line managers who motivate and support their people at the center of engagement and treats their active involvement as a condition for engagement work to succeed.[5]

The most quoted number belongs to Gallup, which attributes about 70% of the variance in team engagement to the manager and calls the manager level the biggest opportunity to improve it.[2] That figure deserves care as vendor research, though it points the same way as the independent work. A polished company program with no manager conversation underneath it is the common way to run a survey and the weakest, because the results stop one level above where the experience is actually made.[1]

Scores stall for a handful of concrete reasons, most of them about which action gets taken rather than whether any action is taken.

Acting on the easy items and avoiding the real driver

Teams often fix the visible, comfortable items and leave the actual driver of low scores untouched. Better snacks and a tidier meeting invite change nothing if the real problem is a manager who does not listen or a workload no one will address. The follow-up review found dissatisfaction persisted wherever the substance of the feedback went unaddressed, whatever peripheral activity took place.[1]

Managers who never hold the conversation

A plan that lives in HR while managers stay silent produces no movement, because the rated conditions are set at the team level. The barriers practitioners name most, execution and follow-through, describe this failure exactly, which is why manager ownership and a little training matter more than the polish of the dashboard.[1]

Too many initiatives, none finished

A dozen simultaneous action items compete for the same scarce attention and tend to stall together. Viva People Science advises against piling on new commitments without retiring old ones, so that a small number of changes actually get finished.[4]

Treating the score as the goal

Chasing the number instead of the experience it stands for tends to corrupt the number. Once a score becomes a target managers are judged on, there is pressure to lobby for higher ratings, time the survey for a favorable moment, or coach teams on how to answer, a workplace version of what social scientist Donald Campbell described: the more a measure drives high-stakes decisions, the more it gets distorted and the less it reflects what it was meant to.[8] A number moved this way rises while the experience stays flat, and the next survey or the turnover data eventually exposes the gap.

Telling a real improvement from a statistical artifact comes down to expecting regression to the mean and comparing like with like across rounds. A rise in a low team's score is the single most easily misread result in the exercise.

Regression to the mean

The team that scored worst last time will usually score better this time even if nothing changed, because extreme results are partly chance and chance does not repeat. Regression to the mean is a statistical phenomenon that can make natural variation in repeated data look like real change, and it should always be considered as a possible cause of an observed movement.[6] It bites hardest exactly where engagement programs aim, at the lowest-scoring teams chosen because they were extreme, so the bounce-back that follows can flatter an intervention that did very little.[7]

Changed questions and changed timing

A score that moves because you reworded an item or shifted the survey date has told you nothing about the workplace. Comparisons hold only when the questions, the scale, and the population stay stable from round to round, and a survey run days after a bonus payout or a layoff reads the mood of that event more than the effect of any plan. Note what else happened in the business during the window before crediting the change to your work.

Attributing honestly

Credit an action only when the improvement is large enough to clear ordinary fluctuation, shows up on the specific items you worked on, and holds across more than one measurement. A single quarter's uptick on the headline number is the weakest evidence that anything real occurred, and treating it as proof is how a program convinces itself it is working.