Theory X and Theory Y: What Your Controls Say About Your Team

A 240-person mid-size Brazilian technology services company (figures in Brazilian reais). To buy an R$ 900 chair, the request goes through three approvals: direct manager, area director, and finance. The average cycle takes eleven days. That same month, an analyst closed an R$ 40 mil deal alone over the phone, because a client called in upset and he handled it on the spot. Nobody approved anything. Everyone thought it was great.

Both things coexist in the same company without anyone blinking. That's because the approval chain wasn't designed around risk: it accumulated. Each layer was born from a specific incident, stuck around, and became policy. What's left is a system that treats whoever asks for a chair as a suspect and whoever gives a discount as an adult.

01The idea: two hypotheses about people #

Douglas McGregor was a professor at MIT when he published The Human Side of Enterprise in 1960. His thesis: every management practice rests on an implicit theory about human nature — and almost no manager knows what their own theory is.

He named the two. Theory X assumes the average person dislikes work, avoids responsibility, and needs to be directed, controlled, and threatened to produce. Theory Y assumes that putting effort into work is as natural as resting, and that, under the right conditions, people seek responsibility and exercise self-control toward goals they're committed to.

The subtle point that gets lost in summaries: McGregor wasn't saying Theory Y means 'being nice' and Theory X means 'being tough.' He was saying these are hypotheses about reality, and that the organization produces evidence to confirm whichever one it adopted. Tight control breeds passive behavior, passive behavior justifies more control, and after a few years the manager has plenty of proof that they were right all along.

McGregor was honest about the limits. He acknowledged that Theory Y requires conditions — a clear goal, available information, real consequences — and that applying it without those conditions is abdication, not delegation.

02Why it still holds up #

The vocabulary changed: today we talk about autonomy, ownership, empowerment. The context of work changed too, becoming distributed and less observable — and that's exactly where the implicit hypothesis shows up hardest. When a manager can't see the person, they have to choose: install monitoring software or agree on deliverables.

What didn't change is the self-fulfilling prophecy mechanism. And what didn't change is that the control system is public: people read into it what the company thinks of them, with far more attention than they pay to leadership's speeches.

03The cost of distrust #

Let's stay with the 240-person company, with 90 managers and coordinators. Explicit assumptions, and everything here is an order-of-magnitude estimate:

The replacement estimate uses the common rule of three to six months of salary, covering recruiting, manager time in interviews, ramp-up, and lost productivity. At R$ 6 mil a month, three months add up to R$ 18 mil — a floor, not a ceiling. If the department that controls the most is also the one that replaces the most people, the correlation deserves at least a conversation.

04Control points versus context points #

Want to see how this looks inside a real operation? Explore the platform.

A practical distinction I use to review processes. Every point where a manager intervenes in a workflow is one of two things:

• Control point — exists to stop something. It blocks the flow until someone authorizes it. That makes sense when the risk is material, irreversible, or regulatory: a payment above a certain amount, access to sensitive data, a contract change.
• Context point — exists to inform someone. It doesn't block anything: it notifies, logs, makes something visible. That makes sense when the manager needs to know, but doesn't need to decide.

Most of the approvals I find in a mid-sized company are context points disguised as control points. The director isn't going to reject the chair. He just wants to know it was bought. Turning that approval into a notification gives nine days back to the cycle without adding any risk.

05Where this gets stuck in practice #

1. Approval levels don't have value tiers. An R$ 200 purchase and an R$ 200 mil purchase follow the same path, which teaches the approver to click without reading.
2. Information doesn't flow without approval. Since the only way a manager finds out about something is by being an approver, they end up approving everything.
3. Whoever has the context doesn't have the authority. The person closest to the problem has to ask permission from someone who knows less about it — and the decision gets worse.

06What data-driven management answers #

Trading control for context only works if the context actually gets delivered. Without visibility, autonomy turns into darkness, and within six months control comes back twice as strong — usually after an incident.

In practice, that means: every decision logged with an author, a date, and a rationale; an exceptions dashboard instead of an approval queue; and a searchable history for when something goes wrong. That's what the platform's people and process modules support, and it's the same logic behind treating the record of work as the source of the metric, instead of questioning the person every time a number moves. For a department-by-department view, the solutions by department help make the idea concrete.

One caveat matters: data about work can be used to give context or to surveil. Same information, two opposite uses. Choosing between them is, literally, choosing between X and Y — and that choice also determines what the company is able to learn and retain.

Nobody reads the values poster on the wall. Everybody reads the approval form.