---
title: "How to build a business operating system | Mallard"
description: "A practical guide to building a business operating system for your scale-up – people, practices and technology, in the right order, without bureaucracy."
url: https://www.mallardhq.com/guides/how-to-build-a-business-operating-system/
modified: 2026-09-16T00:00:00.000Z
---

Business operating systems

# How to build a business operating system that fits.

To build a business operating system, start with people – clear ownership and decision rights – then add a simple, regular rhythm of meetings and playbooks, and only then connect your tools and data around them. Mallard’s advice is to build the minimum that lets the business run without the founder, make it a habit and stop there.

By [Simon Wakeman](https://www.mallardhq.com/our-team/), Principal Consultant & Fractional COO  
16 September 2026 · 8 minute read

Nobody starts a business hoping to build an operating system. The aim is usually simpler: to stop being the person every decision, question and problem comes back to. An operating system is how you get there.

This guide is about the how. The definition – how your people, practices and technology work together as an integrated whole – lives on our page explaining [what a business operating system is](https://www.mallardhq.com/what-is-a-business-operating-system/), so it isn’t repeated here. What follows is the order to build in, what to put in place first and how to stop before it turns into bureaucracy.

## How do you know you’ve outgrown the way you run things?

The clearest sign is that everything still comes back to you. You’ve hired good people, but decisions, problems and questions keep landing on your desk.

Around that, the same symptoms tend to cluster: work duplicated or dropped, capable people who won’t decide, constant messages with no shared context and different teams quoting different numbers. They feel like separate problems. They’re usually one: the business has grown past the informal way it was designed to run. We look at why that happens, and why nobody notices, in [what worked at 12 people stopped working at 25](https://www.mallardhq.com/insights/what-worked-at-12-people-stopped-working-at-25/).

The tempting response is to hire more people or buy another tool. Neither works on its own, because the problem isn’t capacity or software. It’s design.

## Why does the order matter?

Because each part depends on the one before it. People come first, then practices, then technology.

A brilliant meeting rhythm with no clear owners turns into a talking shop. Clear owners with no shared data argue about whose numbers are right. Tools bought before either just automate the confusion. Build in the wrong order and you end up redoing the earlier work anyway.

## How do you get the people part right?

Start by making ownership explicit. Most founder-led businesses run on implied roles that made sense when everyone sat together. As you grow, those gaps and overlaps become where things fall through.

**Name owners for workflows, not just functions.** Your org chart shows who reports to whom. It doesn’t show who owns getting a new client from signed contract to first delivery. Map the handful of workflows closest to revenue and give each one an owner who is accountable for the whole thing end to end, not just their part of it.

**Define decision rights.** For each owner, be clear about which decisions they can make on their own, which they make after asking for advice and which still come to you. When people are told they’re trusted to decide but nobody draws the boundaries, they freeze or they guess. Clear freedom within defined limits is what lets people act.

**Build competence before you expect accountability.** People take ownership of the things they feel equipped to handle. If problems keep landing on you, it’s often because nobody else feels able to solve them yet, not because they don’t care. Coach, show your workings and debrief decisions together, rather than stepping in to fix things yourself.

**Don’t hide in the design.** Sometimes the real issue isn’t the structure, it’s a person who isn’t right for the role or a conversation you’ve been putting off. Redrawing the org chart can feel like progress while avoiding it. If clarifying roles wouldn’t change anything, the problem is probably a conversation.

## What does a good operating rhythm look like?

A good rhythm is regular, predictable and slightly boring. Its job is to replace the context that used to spread by overhearing, and to make sure the right conversations happen at the right time without you calling them.

For most founder-led scale-ups, a workable drumbeat looks something like this:

-   **Weekly:** a leadership meeting with a fixed agenda. Review the same handful of numbers, check progress on priorities and work through the issues that need a decision. Keep it to an hour and protect it.
-   **Monthly:** a deeper look at performance and at how the priorities are going, with time to adjust.
-   **Quarterly:** set a small number of priorities for the next three months, each with a single owner, and review what you learnt from the last quarter.

Alongside the meetings, you need two habits.

**Playbooks for the critical work.** Not a manual for everything, just the few workflows that matter most, written so a capable new hire could follow them. When your best person leaves, these are what stop the process leaving with them, which we cover in [when your best person leaves, so does the process](https://www.mallardhq.com/insights/when-your-best-person-leaves-so-does-the-process/).

**A decision log.** Write down what was decided, why and who owns it. Recording the reasoning, not just the outcome, protects you from forgotten agreements and gives the next person the context they need.

One rule matters more than any agenda: the rhythm runs whether things are going well or badly. When the month gets busy, the weekly meeting is often the first thing to go. That’s exactly when you need it most.

## How should technology fit in?

Technology comes last, and its job is to support the people and practices you’ve already designed rather than define them.

**Fewer, connected tools beat more tools.** Most growing businesses accumulate software one problem at a time. Each tool made sense when it was bought. Together they create five versions of the truth, and people spend their time reconciling them. We look at this pattern in [five tools, five versions of the truth](https://www.mallardhq.com/insights/five-tools-five-versions-of-the-truth/).

**Agree where each answer lives.** For each important number – revenue, pipeline, delivery capacity, cash – decide which system holds the trusted version and who owns it. That one agreement removes a surprising amount of friction.

**Give each owner the visibility they need.** A good dashboard is shaped around someone’s accountabilities, not around everything the system can report.

**Be careful with AI.** AI tools amplify whatever you give them. On top of connected tools and clean data, they’re a real advantage. On top of scattered information and unclear ownership, they’ll just produce confident wrong answers more quickly.

And if you’re considering dedicated software to run your operating system, remember that the software won’t do the changing for you. It can make good habits easier to keep. It can’t create them.

## Should you use EOS, Scaling Up or build your own?

That depends on your business. Off-the-shelf frameworks such as EOS and Scaling Up give you a ready-made structure, a shared vocabulary and plenty of tools, and they suit some businesses well.

The risk is adopting one wholesale and forcing your business to fit the template. Mallard’s view is that your operating system should be designed around how your work actually flows. You can borrow useful pieces – a weekly meeting format, quarterly priorities, a scorecard – without taking on the whole methodology. We explore the trade-offs in more detail in [off-the-peg operating systems](https://www.mallardhq.com/insights/off-the-peg-operating-systems/).

## What is a minimum viable operating system?

It’s the least structure that lets people know what’s expected of them, have what they need to do it and get on with it without you. Build that, then stop building.

For most founder-led scale-ups, the minimum looks like this:

1.  **Named owners** for the three to five workflows that matter most, with clear decision rights.
2.  **A weekly leadership meeting** that always happens, with a fixed agenda.
3.  **Five or six key numbers** that the leadership team looks at every week. Not fifteen.
4.  **A small number of quarterly priorities**, each with one owner.
5.  **One shared place** where priorities, decisions and playbooks live.

That’s it. Past that point, more structure often adds friction rather than clarity. And watch for the moment when designing the system becomes a way of avoiding a harder people problem. The system is scaffolding. It can’t have a difficult conversation for you.

## How do you build it without creating bureaucracy?

Founders who resist operating systems usually fear losing the speed and spirit that got them here. That fear is reasonable. Poorly designed systems do create bureaucracy, but well-designed ones create freedom, because people stop needing to ask.

A few principles keep it light:

-   **Start simple, get consistent, then refine.** Consistency matters far more than sophistication.
-   **Track the vital few.** If nobody would notice a metric disappearing, remove it.
-   **Add one thing, remove one thing.** When you introduce a new meeting or report, retire an old one.
-   **Make room before you make changes.** An overloaded team can’t absorb new habits. Decide what stops first.
-   **Treat structure as temporary.** Review it regularly. Small, frequent adjustments are far less disruptive than one big reorganisation every few years.
-   **Model it from the top.** If the founder skips the weekly meeting or ignores the decision log, everyone learns that the system is optional.

## How do you know it’s working?

Judge it by whether the work got better, not by whether people are using the new templates.

The signs to look for:

-   **Decisions are getting faster**, and fewer of them come to you.
-   **People quote the same numbers** for the same thing.
-   **The rhythm survives a busy month.** Meetings happen even when things are on fire.
-   **New starters get useful faster**, because they can find how things work.
-   **Problems get raised earlier**, before they become crises.
-   **You can step away.** If you can take a fortnight off and the business keeps moving, the system is doing its job.

The best test of all is whether the business has got better at changing. An operating system is never finished. As you grow, parts of it will stop fitting, and a healthy business notices and adjusts without waiting for things to break.

## Where should you start on Monday?

Score yourself against the five-item minimum above, marking each as in place, on paper only or missing. Then work on the gaps in this order:

1.  **Decision rights.** List the decisions that came to you last week and mark the ones someone else could have made with clearer boundaries. That list is the first draft of who decides what.
2.  **One end-to-end owner.** Pick the workflow that matters most commercially – winning work, delivering it or getting paid – and give one person accountability for all of it, including the handoffs.
3.  **The weekly meeting.** Give it a fixed agenda and the short list of numbers from your minimum. Commit to running it for the next eight weeks, whatever happens.

## Should you build it yourself?

Some founders build their operating system themselves, and this guide should help. Others want to know exactly where to start before changing anything, or want an experienced operator alongside them while they build it.

The [_Find_ diagnostic](https://www.mallardhq.com/find-diagnostic-audit/) looks at 262 operational data points across your people, practices and technology in 14 days and ends with a prioritised roadmap of what to build first. An [_Improve_ sprint](https://www.mallardhq.com/improve-sprint/) then puts a senior Mallard operator inside your business for three months to fix the biggest bottleneck end to end, alongside your team, so the capability stays when we leave.

### Some commonly asked questions from founders…

### How do I build a business operating system for my scale-up?

Start with people: make ownership and decision rights explicit for your most important work. Then add a light, regular rhythm of meetings, priorities and playbooks. Finally, connect your tools so information lives in one trusted place. Build the minimum that works, run it consistently and improve it as you grow.

### What’s the difference between a business operating system and EOS or Scaling Up?

EOS and Scaling Up are off-the-shelf frameworks, and they suit some businesses well. Mallard treats a business operating system as the specific way your people, practices and technology work together, so it’s designed around your workflows rather than adopted as a template. You can borrow tools from those frameworks without adopting one wholesale.

### Where should I start: people, practices or technology?

People. Until it’s clear who owns what and who decides what, new meetings have no one to drive them and new tools only speed up the existing muddle. Once ownership is clear, the practices and technology that support it are much easier to design.

### What is a minimum viable operating system?

It’s the least structure that lets people know what’s expected of them and get on with it without the founder. Typically that means named owners for critical workflows, clear decision rights, a weekly leadership rhythm, a handful of key numbers and one place where decisions and priorities live.

### Won’t a business operating system make us bureaucratic?

Only if you over-build it. Bureaucracy comes from tracking too much, adding meetings without removing any and documenting things nobody uses. Keep it minimal, review it regularly and cut anything that isn’t earning its place. A well-designed system should make decisions faster, not slower.

### How long does it take to build a business operating system?

Getting a first working version in place is the quicker part. Making the new habits stick takes longer, and the system is never finished – it should keep evolving as the business grows. Most of the effort goes into running it consistently, not into designing it.

### Do I need software to run a business operating system?

Not at first. Many businesses start with a shared document and a spreadsheet. A dedicated tool can help with consistency once the habits exist, but the tool won’t change behaviour on its own. Get the people and practices working first, then choose software that supports them.

### How do I know if my business operating system is working?

Look at whether decisions are getting faster, whether fewer questions land with the founder and whether the leadership rhythm survives busy weeks. If the founder can step away for a fortnight and the business keeps moving, it’s working.

### Why do business operating systems fail?

Most fail at adoption rather than design. Leaders skip meetings when things get busy, scorecards stop being updated and the system quietly fades. The founder and leadership team have to model it first. If they don’t treat it as non-negotiable, nobody else will.

### How does Mallard help founders build a business operating system?

Mallard usually starts with the _Find_ diagnostic, which scores 12 operational capabilities across people, practices and technology in 14 days and ends with a prioritised roadmap. An _Improve_ sprint then puts a senior Mallard operator inside the business for three months to build the most important parts alongside your team.

## Keep reading

- [What worked at 12 people stopped working at 25.](https://www.mallardhq.com/insights/what-worked-at-12-people-stopped-working-at-25/) – Why the informal habits that ran your business at 12 people break around 25, and how to redesign your people, practices and technology before it hurts.
- [When your best person leaves, so does the process.](https://www.mallardhq.com/insights/when-your-best-person-leaves-so-does-the-process/) – Undocumented knowledge is an easily missed risk in a growing team. How to get the way your business works out of people’s heads and into a shared practice.
- [Five tools, five versions of the truth.](https://www.mallardhq.com/insights/five-tools-five-versions-of-the-truth/) – Your CRM, project tool and finance system all tell a different story. Why piecemeal tools create chaos for scale-ups, and how to fix it in the right order.

### Stop being a bottleneck in your own business

[Book your free 30 minute founder review call](https://www.mallardhq.com/book-a-free-founder-review-call/)
