If I Started an Agency Again, I Would Build the System First

I started my agency in 2014 without much experience.

I kept it going for eight years. Clients continued to come in. From the outside, that might sound like proof that the business was working.

Inside, I was burning out.

At the time, I reached a simple conclusion: agency life was not for me.

After spending a few years in a corporate environment, I see the experience differently. The agency model may not have been the whole problem. I had learned how to do the work and serve clients, but I had not yet learned how to build the operating system around the work.

The business survived because I kept filling the gaps.

If I started an agency again, I would not begin by asking how much more I could personally do. I would begin by designing a business that did not need me to do everything.

Survival can hide a weak operating model

When clients keep coming, it is easy to assume the business is healthy.

Revenue is arriving. The team is busy. Work is being delivered. Problems are being solved.

But demand can hide fragility.

A business may be growing while its leader is still coordinating every moving part, answering every difficult question, rescuing delivery, watching the client relationships, and worrying about where the next project will come from.

The business appears to have a team. Operationally, it still depends on one person.

That was the distinction I did not understand deeply enough at the time. I thought my job was to make sure the work got done. I now think my job should have been to build a system in which the right work could be done consistently, measured properly, and improved without every decision returning to me.

I would create a weekly operating rhythm

One of the first things I would enforce is a weekly update from every team or owner.

Not a long report written for the sake of reporting. A short, consistent update that makes the state of the work visible:

• what was completed;
• what outcome changed;
• what happens next;
• which risks or blockers have appeared;
• which decision or support is needed.

Without this rhythm, alignment depends on memory and informal conversations. A concern can remain invisible until it becomes urgent. Two people can believe they own the same task, while another task has no owner at all. The leader learns about problems only when someone needs a rescue.

A useful weekly update creates accountability without requiring constant supervision. It keeps the team aligned, gives the leader an early signal of drift, and allows help to arrive before the deadline has already been missed.

I would measure business outcomes, not only visible activity

Agency work produces many things that are easy to count.

Tasks completed. Campaigns launched. Pages published. Reports delivered. Traffic gained. Rankings improved.

These signals can matter, but they are not the final reason a client hires an agency.

I would spend much more time defining the business outcome with the client. Are we trying to generate qualified leads, improve conversion, grow revenue, increase retention, or reduce an operating cost? Which parts can our work influence? Which leading indicators tell us whether we are moving in the right direction?

This does not mean claiming that the agency controls every business result. It means refusing to confuse visible activity with value.

If a report shows many completed tasks but cannot help us decide whether the client's business is improving, it is incomplete.

I would give sales clear ownership

Clients coming in is not the same as having a healthy sales system.

A healthy pipeline needs ownership. Someone must know which opportunities are active, where each one stands, what the next action is, how likely it is to close, and whether the work being sold fits the capacity of the delivery team.

If I restarted, I would make sales a defined function rather than allowing it to compete with everything else the leader is doing.

That does not necessarily mean building a large sales team from the beginning. It means giving pipeline health a responsible owner, a regular review, and clear measures. Otherwise, sales receives attention only when the business becomes quiet, then disappears again when delivery becomes busy.

That cycle is exhausting. It also makes planning difficult for the team.

I would stop trying to do everything

My instinct was to be helpful. If something affected the client's result, I wanted to solve it.

The cost was that the boundaries of the service became difficult to see.

I have written about why a system begins with what we are willing to say no to. If every request can enter the business through a different path, the team cannot follow a repeatable process. Exceptions keep returning to the leader because only the leader remembers why each promise was made.

If I started again, I would define the service much more clearly:

• what we solve;
• what the client must provide;
• what our standard process includes;
• what output we are responsible for;
• what sits outside the scope;
• how genuine exceptions are priced and handled.

Being helpful does not require absorbing every adjacent problem. A reliable service needs limits that the team and the client can both understand.

The operating model I would build now

Corporate experience did not teach me that large organisations always work better. It taught me why operating disciplines exist.

The purpose of this structure is not to make a small agency behave like a large corporation.

It is to make the important parts of the business visible before they become emergencies.

The leader's job is to build the system

The hardest change would still be my own role.

Doing the work can feel productive because the output is immediate. Building a system is slower. It requires defining standards, assigning ownership, coaching people, reviewing signals, and fixing the conditions that keep producing the same problem.

Sometimes the leader should enter the work. We may need to understand a failure, establish the standard, or help during a genuine exception.

But if the leader permanently remains the person doing the actual work, the business never develops a stronger way to operate.

This is the same lesson behind letting go as a manager. Delegating tasks is not enough. The team needs context, standards, and real decisions it can own. Otherwise, the work moves but the dependency remains.

My agency survived for eight years. I am proud that it did.

But survival is not the same as building a healthy business.

I once believed the experience proved that an agency was not for me. Now I think it showed me what happens when a leader keeps compensating for the system the business has not yet built.

If I started again, I would still care about doing excellent work for clients.

I would just understand that my most important work as the leader was not to do all of it.

It was to build the system that allowed other people to do it well.