What Is a People Business

Consider four companies that appear, on every surface, to have nothing in common.

The first is a creative agency of thirty people above a coffee roaster, all exposed brick and good light, where the work is ideas: a campaign, a rebrand, a launch film that has to make someone feel something. The second is a law firm on nine floors of a tower, panelled and quiet, where the work is counsel: a merger structured, a dispute defended, a risk read correctly before it becomes a loss. The third is a business process outsourcer with tens of thousands of staff spread across a dozen countries, running the claims handling, the collections, and the back office of banks and insurers who would rather not run those things themselves, where the work is operations at industrial scale. The fourth is a two-person recruitment desk in a serviced office, two phones and a shared spreadsheet, where the work is a match: the right person found and placed for a fee.

Put their annual accounts side by side and they share almost no language. One talks about craft and brand. One about billable hours and partnership. One about seats and service levels and shift patterns. One about placements and time to fill. If you told the agency founder she was in the same business as a forty-thousand-seat outsourcer, she would laugh.

She would be wrong to. Underneath the brick and the panelling and the rows of headsets, these four firms are the same animal. They sell the same kind of thing, wear it out the same way, make and lose money for the same reasons, and are failed by their software for one identical, structural reason. They are people businesses. This part of the book is about them, and this chapter is about what that phrase actually means.

The real question

Now the definition has to bear weight, because the rest of Part V is built on it. What is a people business, precisely? What do these firms share that a manufacturer does not? And why does that shared trait matter enough to earn a name and five chapters of a book?

Here is the definition, stated plainly and meant exactly. A people business is a company whose primary product is the judgement, effort, time, and relationships of its people, rather than a manufactured good or a self-serve piece of software. That is the whole of it. There is no product on a shelf. There is nothing a client can pick up, inspect, and compare against a rival's before buying. What is being sold is capacity and judgement: a promise that a group of capable people will apply themselves to your problem and produce something good. The invoice, when it comes, is for services rendered, which is the accountant's polite way of saying: for what our people decided and did.

Say it that way and the four firms in the cold open stop looking like strangers. Every one of them sells the applied judgement of its people. Everything else, the brick, the panelling, the headsets, the methodology decks, is set dressing around that single fact.

It helps to say what a people business is not, because the boundary sharpens the definition. A pure software company is not primarily a people business, because its value is embodied in a reusable product that exists apart from the people who built it and keeps earning while they sleep. A manufacturer is not one, even a manufacturer that prides itself on craftsmanship, because the value it delivers is embodied in an object the buyer can hold. But few large organisations are pure anything. Many software companies contain substantial people-business units, in implementation, consulting, customer success, and managed services. So the category is best treated not as a box a whole company falls into but as a test you apply to a business, or to a unit inside one. When you strip everything else away, is what remains for sale the judgement of people, with no manufactured good and no reusable product standing in front of it? Where the answer is yes, you are looking at a people business, whatever the parent calls itself.

The family, member by member

The phrase covers more ground than most people realise. These firms do not think of themselves as one category. They should.

Agencies, creative, media, digital, and public relations, are the most visible branch. Whatever the discipline, an agency sells taste and effort: ideas and the craft to execute them, the judgement of where a message should go and what attention is worth, relationships and timing judged on outcomes it influences but never fully controls. The work cannot quite be scheduled and the client needs it on Thursday, and the margin is always one over-servicing decision away from disappearing.

Consultancies sell judgement more nakedly than anyone: the quality of thinking about a hard question, the ability to change how an organisation works, expertise a client cannot justify keeping permanently on staff. Strategy, management, and specialist firms differ in subject, but each exists on a single claim, that its judgement is worth more than its cost, and its economics are the daily test of whether the claim holds.

Professional services are the old, credentialed branch: legal, accounting and audit, engineering, architecture, surveying, advisory. A law firm sells counsel, an audit firm sells assurance and the judgement behind the signature, an engineering practice sells designs that must be right because gravity does not negotiate. What they share with a two-person creative shop, despite the gulf in formality, is that the asset is expertise held in people's heads, rented by the hour or the engagement, and worth exactly what the market trusts the name on the door to be.

Staffing and recruitment firms sell the match itself: the right person for the right seat, found faster and more reliably than a client could manage alone. At the specialist and retained end, the product is close to pure judgement, knowing which candidate will actually thrive and which will resign in ninety days, on trust built with clients and candidates over years. At the high-volume end it is judgement wrapped in serious operational machinery: sourcing at scale, compliance, fulfilment, payroll, and the logistics of placing many people quickly. Across the range the core is the same, a stream of repeated decisions about fit, availability, timing, risk, and trust, supported by the operations needed to act on them at speed.

Business process outsourcers, the BPOs, are the largest and least examined member of the family, a vast industry that runs other companies' operations for them. When you make a claim on an insurance policy, chase a late delivery, dispute a charge, or apply for something and speak to someone who helps you, there is a fair chance you are talking to a BPO running that process on behalf of a brand whose name is on the letterhead but not on the payslip. They handle finance and accounting, customer operations, collections, claims, onboarding, content moderation, and the back office of whole industries at scale, across many clients at once, each client with its own contract, its own promised service levels, its own quirks. The BPO sells operational capacity and the judgement to run someone else's process better and more cheaply than the client could themselves. It is a people business measured in tens of thousands of people, and its margins are thin enough that a small error in how work is estimated, staffed, or delivered is the whole difference between a profitable programme and a loss.

Contact centres are close cousins to the BPO, and often the same firm, but they are worth naming on their own because they concentrate the whole problem into a single visible tension. A contact centre sells resolved interactions at a promised quality and cost. Every day is a live negotiation between utilisation and service: staff too lean and the queue grows and the client suffers, staff too rich and the margin evaporates. Plenty shapes the economics, forecasting accuracy, scheduling, channel mix, technology, and process design among them. But the input that coordinates all of those, and that no system captures, is judgement: the thousands of small human calls a day, how to handle this caller, when to escalate, when to hold a policy and when to bend it. The forecasting and the scheduling can be bought. The judgement exercised inside them is the part that leaks away unrecorded.

Managed service providers round out the family: firms that run a client's IT, infrastructure, or security for a recurring fee. An MSP sells the promise that something critical will simply keep working, and its economics turn on the gap between the flat fee and the unpredictable effort each client demands, quiet and profitable one month, a fortnight of firefighting the next on the same fee.

Most real firms are more than one of these at once, which is part of why the category goes under-named. A single group may run an agency, a consultancy, and a managed service under one roof. A law firm looks nothing like a staffing agency until you notice that both are renting out expert time by the hour. The labels blur at the edges. The structure underneath them does not, and it is the structure that matters.

In aggregate these firms are enormous, a huge share of the employment and the output of every developed economy, and yet they are strangely absent from the way we talk about business, which is dominated by companies that make products and companies that make software. The people business is hiding in plain sight, at scale, and under-served by the tools it has been handed.

What they all share

They look nothing alike. They are, structurally, almost identical. Five traits run through every one of them, and every trait is a consequence of the same root fact: the product is judgement, not a manufactured thing.

They sell capacity and judgement, not a finished, inspectable product. A manufacturer sells an object you can hold, test, and return. A software company sells something you can trial before you buy. A people business sells a promise: that capable people will apply themselves well to your problem. The buyer cannot inspect the product in advance, because the product does not exist yet. Trust does the work that inspection does elsewhere. This is why trust is not an abstraction for these firms but the ground they stand on.

They staff against promises made before the work is understood. The sale comes first. The scope, the price, and the deadline are agreed at the moment the firm knows least about the work, and then people are assigned to honour a promise that was made before anyone had really looked at the problem. A factory designs the product, then prices it. A people business prices the product, then discovers it. Every engagement is, to some degree, a bet placed before the cards are turned over, and the whole apparatus of delivery is the work of making that bet come good.

Their margin lives in the gap between what was sold and what it costs to deliver. This is the sentence to keep. A people business does not make its money at the point of sale. It makes or loses it in the distance between the price agreed and the true cost of delivery, and that distance is composed entirely of decisions. Who gets staffed on the work, at what grade. Whether the extra request is absorbed or charged. Whether an escalation is caught early or late. Whether the plan is real or held together by one person's nights and weekends. Each of these is a small choice, and the margin is their sum. The gap between the deal and the done, the gap this book opened on, is not a metaphor. It is where the money is.

Their primary asset walks out of the door every evening, and can resign. A factory's value sits in its plant and its inventory, which stay behind and belong to the firm. A people business keeps its most valuable assets, its institutional knowledge and its client relationships, inside the heads of people who go home at six and are free to leave for a competitor at any time. The senior partner who is the real reason three clients stay is not on the balance sheet, and cannot be put there. When someone leaves, they do not just cost a replacement; they take a piece of the product with them, and often a piece of the memory too.

Their economics are driven by utilisation and the quality of judgement, not by units shipped. A manufacturer improves by making more units at lower cost per unit. A people business has no units. Its two real levers are how fully its people are engaged on paid work (utilisation) and how good their decisions are (judgement), and the two are in permanent tension. Push utilisation too hard and judgement suffers, because tired people staffed on work they do not fit make the small bad calls that leak margin. Protect judgement with no regard for utilisation and the firm cannot pay its people. Running a people business well is the daily act of holding those two in balance, and there is no meter on the wall that shows you either one the way a production line shows you output.

Five traits, one animal. That is a people business, and once you can see the shape you cannot unsee it, whether it is thirty people over a coffee roaster or forty thousand across a dozen countries.

For all that shared structure, these firms do not make their money in the same proportions, and it is worth naming the range, because a category that treats a two-person search firm and a forty-thousand-seat outsourcer as identical in every respect is a category on its way to meaning nothing. People businesses sit on a spectrum. At one end are the judgement-intensive firms, where the client is buying scarce expert judgement: law, strategy, executive search, specialist advisory. In the middle are the craft-and-delivery firms, where judgement is expressed through skilled making and execution: agencies, engineering, implementation consultancies. At the other end are the capacity-and-process firms, where the client is buying reliable human capacity at scale: staffing, contact centres, BPOs, managed services. What unites the whole spectrum, and earns it a single name, is that the value must still be delivered through people, repeatedly, after the sale, so the decisions that shape that delivery are where the firm is made or unmade.

Why the software never fit

The enterprise software these firms run on was built for the other kind of company, and it shows, every day, in a hundred small frictions.

The ERP was built to coordinate the making and moving of goods: bills of materials, stock, orders, the flow of a physical thing through a factory and out to a customer. The CRM was built around the sales pipeline, a funnel of deals moving toward a close. Both are triumphs of engineering, and both encode a worldview in which the valuable thing is a product or a transaction, something settled, countable, and inspectable. That worldview is correct for a manufacturer and a distributor. It is quietly wrong for a firm whose product is a promise and whose value is a stream of judgements.

So the people business is handed tools built to answer questions it does not have, and unable to answer the questions it does. Ask an ERP built for inventory to hold an engagement, a promise to a client that must be staffed, delivered, and defended over months, and it has no natural place to put it. Ask a CRM built for the pipeline what happened after the deal closed, and it loses interest, because in its worldview the close was the finish line, when for a people business the close is the starting gun.

It would be too glib, though, to say the software industry simply ignored these firms. It did not. A whole category of service-specific tools exists and has for years: professional services automation, resource management and workforce planning, practice management, project accounting, field service and contact-centre platforms, and the services modules of the big suites. Many are good at what they do. So the honest version of the problem is not that no one built for services. It is that each of these systems manages a fragment of the people-business model, the resourcing, or the time and billing, or the project plan, or the pipeline, and none of them governs the whole loop from promise to staffing to delivery to margin to learning, or holds the decisions that connect those stages. That is a subtler gap than "the software never fit," and it is the real one.

Picture the Monday staffing meeting that almost every firm of this kind holds in some form. A dozen people gather, in a room or on a call, to decide who works on what this week. The information they need is scattered across the systems: the CRM knows what is about to close, the finance system knows what is over budget, the resourcing tool knows who is nominally free, the project tool knows what is slipping, and none of them agrees with the others or, quite, with the truth. So the meeting runs instead on a spreadsheet someone rebuilt by hand over the weekend, and on the memory in the room. It produces a set of consequential decisions, this person onto that account, this engagement given another week, this hire brought forward, that are made, acted on, and then never recorded anywhere as decisions. By Friday the spreadsheet is stale. By the following Monday the reasoning behind last week's calls is gone, and the meeting begins again from a standing start.

The result is felt not as a strategic problem but as a daily texture of workarounds. The real operating model of most people businesses does not live in the expensive systems at all. It lives in spreadsheets: the staffing plan, the true margin tracker, the pipeline the leadership actually believes, the list of accounts quietly at risk, all maintained by hand because no system holds them properly. And it lives, above all, in heroics: the account manager holding three engagements together with personal effort, the delivery lead who is the only person who really understands the plan, the partner who carries the client relationship entirely in her head. The workaround, the spreadsheet, and the heroic are not signs of a badly run firm. They are the scar tissue that forms wherever the software was built for a different species and the people had to route around it to get the work done.

The natural home of Decision Intelligence

Recall what Decision Intelligence is: the practice of treating decisions as things worth keeping, capturing them, connecting them to the evidence they rested on, governing them, and learning from how they turn out. In any company that is valuable. In a people business it is close to existential. For these firms, success is not a product that happens to require good decisions along the way. Success literally is a stream of decisions: who does what work, for whom, at what price, at what quality, staffed how, escalated when, renewed or released, fought for or let go. There is no manufactured object standing between the judgement and the result. The judgement is the object. The decisions are the product.

That is why the absence of a decision layer, which is a general problem for every company, becomes an acute one here, and why it strikes in the two places a people business can least afford to be struck. It hits margin, because the margin is the sum of a thousand small decisions in the gap between sold and delivered, and when those decisions are never captured, the margin leaks through them unseen and unlearned. And it hits trust, because trust is a people business's entire basis for existing, the thing the client buys in place of an inspectable product, and trust is built only by a consistent record of good decisions over time, which is precisely the record these firms cannot currently produce. Every company pays for poor decision memory, and some pay catastrophically: a manufacturer's bad call can wreck a supply chain or trigger a recall. The difference in a people business is directness and speed. Because judgement is delivered straight through the service, with no product in between, a missing decision layer shows up almost at once in the two assets that are the whole company: its margin and its good name.

There is a compounding buried in this that is easy to miss. Because the decisions are the product, a people business that learns to keep them does not merely avoid repeating a mistake; it slowly turns the private judgement of its best people into an asset the whole firm can draw on. The pricing instinct that currently lives in one partner's head becomes a record the next person can reason from. The reason a difficult client was worth keeping, or worth releasing, survives the departure of the person who made the call. In a firm whose entire value is judgement, keeping the judgement is how the firm gets better at the one thing it sells. That is why the discipline pays back here with an interest that a product company, sheltered behind its product, will never quite feel.

It is the place the discipline was, in a sense, always aimed. Where the product is judgement, a practice for keeping and improving judgement is not an optimisation. It is the missing core.

"But surely every business is a people business now"

Here is the fair objection, and it deserves a straight answer. Every business is a people business now, you might say. Software firms are famously all about talent. Manufacturers compete on the skill of their engineers. Banks are, when you look closely, enormous knowledge organisations. If judgement and people matter everywhere, then a category that includes everyone describes nothing.

The objection is half right, and the half it gets right is worth conceding generously. It is true that the discipline of this book generalises. Every organisation makes decisions, every organisation would be better for keeping them, and the argument for a decision layer does not stop at the edge of the professional services sector. It is also true that the modern software company, or the modern manufacturer, has real people-business characteristics, and would recognise some of the five traits in itself.

But there is a difference between a business that has people and a business whose product is its people, and it is the difference between a trait and an essence. A software company has brilliant people, but it also has a product: a piece of software that exists, can be inspected, runs whether or not its author is at their desk today, and can be sold a millionth time at almost no extra cost. That product is not immunity from bad judgement, product companies are wrecked by roadmap errors, safety failures, mispricing, and bad bets on where to spend their capital. But it is a buffer. A product embodies past judgement in a reusable asset, so a good decision keeps paying out and a bad one can often be absorbed, corrected, and shipped over. The people business has no such buffer. The judgement is exposed, directly and at once, as margin and as trust.

That is why the pure cases feel the problem first and most acutely. They are the firms with nowhere to hide. When your entire product is the applied judgement of your people, the failure to keep and improve that judgement is not one inefficiency among many. It is a hole in the bottom of the boat, and you are already taking on water. The people business is where the case for a decision layer stops being an argument and becomes a matter of survival.

The bridge

So the people business is a real and enormous species, and it is the place the missing decision layer is felt most sharply. We have named the two wounds: margin and trust. The rest of this part goes into each of them.

It goes to margin first, because margin is where the mismatch does its quietest and most expensive damage. Every one of the five traits pointed at the same place: the gap between what was sold and what it costs to deliver, filled with small, unrecorded decisions, is where a people business makes its money or loses it. That gap does not fail loudly. It does not throw an alarm. It leaks, one absorbed request and one mis-staffed engagement at a time, in a way no monthly report catches until the money is already gone. The next chapter is about that leak: where it happens, why it is nearly invisible while it can still be stopped, and how a firm might finally learn to see it in time.