Companies have never spent more effort on culture than they do right now. There are Chief Culture Officers, values off-sites, belonging councils, and engagement platforms that report a number every quarter. An entire profession exists to make work feel better.

The results have moved in the opposite direction. Gallup's latest global data puts employee engagement at a five-year low, with roughly one in five employees engaged and a second consecutive year of decline. Manager engagement, which is supposed to carry the whole system, has fallen even faster, and the estimated cost runs into the trillions in lost productivity. In any other part of the business, rising investment against falling output would trigger a hard question about the model itself. Culture seems to be the one budget line where we keep blaming the execution instead.

So ask the harder question. What if culture building, as most companies practice it, was aimed at the wrong target from the beginning? Three forces have made that possibility difficult to keep ignoring.

The first is credibility. In 2025, U.S. employers announced more than 1.1 million layoffs, one of only a handful of years above that line since 1993, and the cutting has settled into what recruiters now call rolling layoffs, a steady drip rather than a single event. Employees increasingly describe their leaders as misaligned, even hypocritical. You cannot run a values workshop on Tuesday and a quiet headcount reduction on Thursday and expect anyone to believe the poster on the wall. Culture programs rarely fail because the words are wrong. They fail because people can see the distance between the words and the decisions.

The second is that most culture work treats symptoms. When MIT Sloan researchers analyzed what actually drove the Great Resignation, a toxic culture predicted attrition roughly ten times more strongly than pay. But look at what toxic meant in the data: people feeling disrespected, leaders tolerating underperformance, good work going unrecognized. None of that is repaired by a perks budget. Those are operating problems, and the perks were never going to reach them.

The third is AI. In a 2025 study, researchers from Harvard, Columbia, and Yeshiva University found that adding automation to teams often decreased overall performance, increased coordination failures, and eroded trust, with the worst effects falling on less experienced teams. AI accelerates the individual while straining the connective tissue between individuals. The bottleneck has moved from how much each person can produce to whether the team can decide, hand off, and align, which is exactly the layer a culture program never touches.

All of which leads to a conclusion most culture budgets are built to avoid: you cannot build culture directly. You never could. Culture is a lagging indicator. It is what accumulates when a team makes decisions, ships work, recognizes results, and manages its energy the same way for long enough that it hardens into how we do things here. Every culture you admire sits downstream of a team that performed first. The language came afterwards.

Most companies run that sequence in reverse. They try to manufacture the feeling and hope the performance follows. The better order is the honest one: build the performance, meaning the behaviors, the standards, and the accurate recognition of who is actually moving the needle, and let the culture arrive as the byproduct. Culture built that way survives a bad quarter, because it was formed in the work rather than announced at an off-site.

Here is a test worth running on your own organization. If the culture budget went to zero tomorrow, what would actually change about how your best team decides, ships, and recognizes work? If the honest answer is very little, you already know where your real culture lives, and it was never in the budget.

o o o o o o o o 

Performance First is a KOhesion series.