Handing Over Responsibility Without Losing Control

Why responsibility so often wanders back to the owner

An owner hands customer service over to an experienced employee. In the first weeks he asks daily, reads every email, and corrects wording. After a month the employee says she does not feel responsible. He withdraws, hears nothing for three months, and learns from an annoyed major customer that complaints have been sitting for weeks. After that, he takes over again himself.

Both mistakes have the same cause. There was no agreement on how he would recognize that things were going well without looking. The guide delegating tasks with a clear handover describes how individual tasks are handed over. This guide goes one level higher, to entire areas, and asks: How does an owner hand over responsibility permanently without losing control, and what must be agreed for this?

To answer it, I first clarify the terms and present a thinking model: Robert Simons' four levers of control. This is followed by findings, a weighing between supervision and trust, and finally the responsibility agreement with an example.

Responsibility, supervision, agreement: the terms

Delegating means transferring an outcome to a person and leaving them the decisions needed along the way. Overall responsibility toward customers and the business remains with the owner. This is how the guide on delegating defines it. Handing over responsibility here means doing this permanently for an entire area, such as purchasing, customer service, or a branch, and not just for a single task.

Control means, as in the essay trust in the team, all arrangements that restrict latitude or monitor behavior. For this guide I distinguish two forms. Supervision checks behavior during the work: reading along, asking, approving. An agreement sets out in advance what someone is responsible for, which limits apply, which numbers show success, and when it will be discussed. Afterwards the owner checks results and exceptions, not every step.

The thinking model: Simons' four levers of control

In his 1995 book “Levers of Control,” Harvard professor Robert Simons described how managers can give freedom and keep control at the same time. His starting point is a tension: employees should seize opportunities and act independently, but the business must remain predictable. Simons does not resolve this with a single tool but with four levers that work together.

The first two levers provide a frame. The belief system says where the journey is going, the boundary system where the path ends. The last two levers provide information. The diagnostic system measures results against target values and only speaks up when there are deviations. It relieves the owner because he does not have to look as long as the numbers are within range. The interactive system is the opposite: a regular conversation about the questions nobody has a ready answer to. This is where the business learns.

An everyday analogy shows the interplay. Parents who lend their child the car for the first time do not ride along. They say what the car is for, name a few rules that are not negotiable, agree on a time to return and a full tank, and the next morning they ask how it went. These are the four levers in a family.

Whoever hands over gives up watching, not knowing.

What studies show about the four levers

Widener (2007) tested the model with 122 chief financial officers. The four systems were related to each other and complemented rather than replaced one another. At the same time, the study showed the costs and benefits: control systems consume management attention, and they promote learning in the company. Kruis, Speklé, and Widener (2016) surveyed 217 managers and found four different patterns of how the levers are kept in balance. So there is no single right mix.

For small businesses, a new study from Denmark is revealing. Anees, Sponem, and Di Berardino (2026) surveyed 112 small and medium-sized enterprises. They examined the interactive lever, that is, structured conversations in which management uses budgets and metrics to develop a shared understanding with employees across all levels. Where these conversations took place frequently, the relationship between the CEO's creativity and the company's creativity was stronger. In family firms this effect was more pronounced than in others.

The authors explain this through two pathways. In the conversations, employees receive the information and decision-making authority they need for their own initiatives. And in recurring meetings they see how the owner asks questions and tackles problems, and adopt this attitude. It is a one-time survey; causes cannot be reliably derived from it.

How employees respond to transferred responsibility is shown by the meta-analysis of Kim, Beehr, and Prewett (2018), which the guide on delegating presents: across 55 samples, a clear relationship with the evaluation of the leader (0.59) and a weaker but positive one with behavior and performance (0.31).

Supervision or trust? Two views

The first view relies on supervision. It rests on principal-agent theory, described in the essay principal-agent theory in the business: the employee knows more about his work than the owner and does not automatically pursue the same goals. Whoever does not look learns of problems too late. Especially with customers, money, and liability, this is a serious argument.

The second view relies on trust. It rests on stewardship theory, which the same essay presents as the counter-model: whoever identifies with the business acts in its interest, and constant checking damages precisely this identification. The essay trust in the team shows why control often replaces trust and rarely improves it.

I consider both views incomplete, because they equate control with supervision. Simons' model shows a third way: the owner gives up watching, but not information. Boundaries take away the risk that principal-agent theory emphasizes. Metrics with thresholds ensure that problems are reported without anyone having to ask. And the regular conversation creates the identification on which stewardship theory relies. Trust is then no longer a hope but the result of a good agreement.

The responsibility agreement on one page

From the model I derive a tool that I call the responsibility agreement. It applies the four levers to one area and fits on a single page. It is written together with the responsible person, not for them.

1. Purpose in one sentence

   What does the area exist for, and how does a customer notice that it is well run? The sentence helps with all decisions no rule anticipates.

2. Limits with value thresholds

   What must never happen, and what only after consultation? A value threshold is an amount up to which a decision right applies, as described in the guide setting decision paths. Three to five limits are enough.

3. Two to four metrics with thresholds

   Which numbers show whether the area is running, and from which value does the responsible person report on their own? How such numbers fit on one page is shown in a one-page KPI dashboard.

4. Fixed meeting date

   Once a month, 30 minutes about what has changed and what nobody knows for sure. Not a report on what has been done, but a conversation about questions.

5. Trial period for the agreement

   After three months, you check together whether limits and thresholds fit. Only then does it apply indefinitely.

How much time this saves is shown by a worked example for an owner who has handed over three areas. With supervision, he spends the time reading along and asking. With an agreement, he reads metrics, holds the monthly conversations, and deals with reported exceptions.

When the owner should step in again

An agreement does not mean never stepping in again. It means deciding in advance when. Three occasions justify it: a limit was exceeded without consultation, a metric has been beyond the threshold for several weeks without a plan, or the question concerns the whole business.

Even then, a graduated response helps. The guide on delegating describes four levels of latitude, from executing through proposing and deciding and reporting to deciding. Instead of taking an area back entirely, the owner can temporarily go back one level for a particular type of decision and say so explicitly. Whoever silently starts deciding again makes the whole agreement worthless.

Why this letting go is particularly hard for founders is described in the essay from founder to entrepreneur.

What you can do this week

• Choose an area where you still ask too often or know too little today.
• Write down with the responsible person the purpose in one sentence and three to five limits.
• Set two to four metrics with thresholds at which they report on their own.
• Schedule a monthly 30-minute conversation for the next three months.
• During this time, stop reading along in that area.

Conclusion: How do you hand over responsibility without losing control?

The opening question was
an owner hands over responsibility permanently without losing control, and what must be agreed for this. The answer lies in the distinction between supervision and steering. Supervision costs more time with every area handed over and teaches employees to seek cover for decisions. Steering through Simons' four levers, that is, purpose, boundaries, metrics, and regular conversation, gives the owner the information he needs without having to look.

The studies support this: the levers complement each other, and particularly in small family businesses, regular conversation about numbers strengthens what employees contribute themselves. So it is not whoever hands over who loses control. Control is lost by whoever hands over without saying in advance how they will recognize that things are going well.