The recognition gap: the highest-ROI lever teams ignore

A staff data engineer I know spent three weeks untangling a silent data-quality bug — a timezone mismatch that had been quietly corrupting a revenue dashboard for months. She found it, fixed it, backfilled the bad rows, and wrote a postmortem so other teams would not repeat it. The dashboard went green. Nobody said anything. The dashboard had simply stopped being wrong, and a thing that is no longer wrong does not show up in anyone’s notifications.

A week later, in a one-on-one, her manager mentioned in passing that he “gives the team plenty of recognition.” She nodded. Privately, she could not remember the last time he had told her a specific thing she did well.

Neither of them was lying. That small, two-person disagreement — the manager certain he praises constantly, the report unable to recall it — is one of the most reliable findings in modern workplace research, and it points at the cheapest high-return lever most teams never pull.

Two people, one team, two different jobs

Gallup, which has been measuring the inside of the workplace for decades across more than a hundred countries, keeps surfacing a particular kind of mismatch. Ask managers whether they give meaningful feedback every week and about half say yes. Ask the people who report to those same managers whether they actually receive it, and only about one in five agree. Recognition tells nearly the identical story: managers believe they hand it out routinely; their reports experience a near-silence.

That is not a rounding error. It is a roughly two-and-a-half-to-one gap between intention and reception — two people sitting in the same one-on-one walking away with opposite accounts of what just happened.

It is worth pausing on why the gap is so wide and so stable, because the explanation is not that managers are careless or that employees are needy. The two sides are doing genuinely different jobs. The manager remembers the act of giving — the encouraging Slack reaction, the “nice work” in standup, the kind line in a review. The report only registers recognition that was specific, timely, and unmistakably for them. A thumbs-up emoji counts in the giver’s memory and evaporates in the receiver’s. The accounting is asymmetric, so even an honest manager overestimates, and even a fairly supported employee can feel unseen.

Why this hits data and engineering teams harder

Every team has this gap. Software and data teams have it worse, for a structural reason: a lot of our best work is, by design, invisible.

When a feature ships, there is a demo and a launch post and a screenshot in the channel. But the engineer who deletes four hundred lines of dead code, the analyst who catches that a join was silently dropping ten percent of rows, the platform person who tunes a query so it stops paging people at 3 a.m. — their wins look like nothing happening. Reliability is the absence of incidents. Good data is the absence of a wrong number. Refactoring is the absence of future pain. The reward for excellent infrastructure work is that no one ever has to think about it, which is also, accidentally, the reward structure for being forgotten.

Remote and hybrid work thins the channel further. The hallway moment where a senior engineer says “that migration you ran was clean” does not happen on its own over async tools; someone has to decide to type it. And AI coding assistants are quietly raising the bar for what counts as noteworthy — when a model can scaffold a service in an afternoon, the human judgment that made it the right service, safely rolled out, becomes even easier to overlook precisely because it is harder to see. The work that is most worth recognising is increasingly the work that leaves the least visible trace.

So the perception gap is not just a soft-skills footnote for us. It is the default outcome of how engineering value is distributed: concentrated in prevented disasters and quiet correctness, exactly the things that generate no artefact to point at.

The cheapest lever with the best numbers

Here is what makes the gap worth caring about rather than just noting. Recognition and clear expectations are nearly free, and the return on them is unusually large for something that costs no budget.

Gallup’s numbers put rough figures on it. People who receive regular, meaningful recognition report something like 22 percent higher satisfaction with their jobs. Setting clear expectations — telling people what good looks like, which is the other half of feedback — is associated with roughly 30 percent higher engagement. And in a finding that should be printed on every manager’s monitor: about 65 percent of employees say they would work harder if they felt their efforts were genuinely noticed. Not paid more. Noticed.

Compare that to where wellbeing money usually goes. An Oxford study led by William Fleming, covering more than 46,000 UK workers across some ninety different interventions, found that the popular individual-level perks — mindfulness sessions, resilience training, wellbeing and sleep apps, one-off coaching — showed little to no measurable benefit, and a couple even trended slightly negative. Recognition was not one of the things that failed. It belongs to the category the same body of research keeps pointing back to: changing how the work and the management around it actually feel, rather than bolting a perk onto an unchanged job. A meditation-app licence has a per-seat invoice; telling someone exactly why their fix mattered costs a sentence.

There is a heavier number underneath all of this, too. Roughly 69 percent of employees say their manager affects their mental health more than their doctor or their therapist — on par with their partner. That statistic gets quoted to scare managers, but the honest reading is more hopeful: it means the person best positioned to improve how their team feels is not a benefits vendor or an EAP hotline. It is the lead who notices the timezone-bug fix and says so.

What good feedback actually looks like

If the gap is a delivery problem, not a budget problem, then closing it is a craft you can practise. A few things separate recognition that lands from recognition that evaporates.

Be specific enough that it could only be about them. “Great job this sprint” is the emoji of feedback — pleasant, instantly forgotten, indistinguishable from what you would say to anyone. “The way you instrumented that pipeline so we caught the schema drift before it hit the dashboard saved us a Monday fire drill” is recognition the receiver will actually log, because no one could have said it about someone else. Specificity is what converts a giver’s intention into a receiver’s memory.

Make it close to the event. Praise has a short half-life. Recognition delivered the day of the fix is worth several times the same words in a quarterly review, when the moment is cold and the praise reads as a formality. The point of a weekly one-on-one is not the meeting; it is that it keeps the distance between good thing happened and good thing acknowledged short.

Separate recognition from the performance-review machine. When the only feedback people get arrives bundled with ratings, calibration, and compensation, every kind word gets read as a maneuver. Day-to-day recognition has to live outside that system to be believed — a genuine “this was sharp,” detached from anything that feeds a score.

Treat clear expectations as half the gift. Feedback is not only praise; it is telling people what good looks like before they are judged against it. On data teams especially — where “done” can mean anything from the query returns to it is tested, documented, monitored, and someone else can own it — ambiguity is a quiet tax. Naming the standard up front is itself a form of respect, and it is the lever Gallup ties to the biggest engagement lift.

And a caveat worth saying out loud, because these are survey numbers from different instruments with different definitions: treat the magnitudes as directional. A “22 percent lift” is not a guarantee you will see 22 percent of anything on your team. The robust, repeated signal is the direction and the cost — recognition and clarity reliably help, and they are close to free. That is a strong enough bet to act on without pretending the decimals are precise.

The move

If you lead a data or engineering team, you do not need a program, a platform, or a budget line to close this. You need to override a default. The default is that the timezone-bug fix turns the dashboard green and disappears, and three weeks of careful work generates exactly zero acknowledgement because it produced no artefact. Overriding it costs one specific, timely sentence, delivered outside the review cycle, to the person who would otherwise assume nobody saw.

And if you are the engineer whose best work keeps vanishing into the pipeline: some of this is yours to fix from below. Make the invisible legible. Write the postmortem, post the before-and-after, say in standup what the silent fix prevented — not to boast, but because you are doing your manager’s noticing for them, closing the gap from the other side.

The cheapest, highest-return lever in the whole wellbeing conversation is not an app or a stipend. It is one person telling another, precisely and soon, that the work mattered. The strange thing is not that it works. The strange thing is how reliably we forget to do it.