I have watched, over twenty-five years, the job I do get quietly chopped into pieces. When I started, a salesperson did the whole thing: found the account, earned the meeting, ran the discovery, built the case, closed the deal, and kept the customer alive afterward. One person, one relationship, start to finish. Today that same arc is divided among four or five specialists who may never speak to each other, each owning a slice, each measured on it. This is progress, I am told, and it is certainly efficient. Whether it is progress depends on a question almost no one asks first, and Adam Smith asked it in 1776.
Smith opened The Wealth of Nations with a pin factory. One worker doing every step of pin-making produces almost nothing; ten workers each doing one step, one drawing the wire, one straightening, one cutting, one pointing, produce thousands. Division of labor multiplies output, and it is the founding idea of the modern economy. It built the factory. Eventually it came for the sales floor.
The blueprint has a name
Aaron Ross's Predictable Revenue, published in 2011 and often called the sales bible of Silicon Valley, took Smith's factory and rebuilt it for selling. Stop asking one person to do everything. Split the seller into a line of stations: the development rep who only prospects, the account executive who only closes, the solutions engineer who only handles technical proof, the customer success manager who only renews. Each station repeats one narrow operation until it is fast and cheap and measurable. Output per head climbs. The org chart becomes a pin factory, and for a decade it was sold as the one correct way to build a revenue engine.
The condition nobody checks
Division of labor only works when the task can actually be divided, when it cuts into steps that hand off cleanly, each worker receiving a finished input and passing a finished output. Herbert Simon called such systems nearly decomposable: strong connections inside each part, weak connections across the seams, so you can cut at the seams and lose almost nothing. A pin does not care that five hands made it. The wire remembers nothing. Each station gets a complete thing and passes a complete thing. The joints are clean.
Some selling is exactly like that. Transactional, high-velocity sales, a defined product, a short cycle, a buyer whose need is legible on contact, is nearly decomposable, and there the assembly line is correct. Ross is right. The specialists genuinely beat the generalist, because the handoffs are cheap and the gains from repetition are real. If that is the business you are in, industrialize it, and do not let me talk you out of it.
Where the line severs
Complex enterprise selling is not like that, and here the assembly line quietly severs the product at every handoff. What gets built across a nine-month enterprise deal is one continuous thing: trust, context, and a compressed understanding of what the buyer actually needs underneath what they said they need. That understanding lives in a relationship and cannot be serialized onto a lead-routing rule. Every handoff cuts a link that was carrying full weight. The prospecting-to-closing handoff drops the tacit half of why the buyer engaged, the part that never fit the form field. The closing-to-renewal handoff drops the promises and the read of the room that won the deal, so the customer re-explains themselves to a stranger the week after signing. In a decomposable task the handoff is free. In this one the handoff is a tax, paid in lost context, and paid again at every station.
Smith saw the other cost too
He put it in the same book, and everyone forgot. Hundreds of pages after the pin factory, in Book V, Smith warns that a man who spends his life performing a few simple operations becomes as stupid as it is possible for a human creature to become. Read that on a modern development-rep bench, a role stripped to dial, log, book, repeat, and the eighteen-month burnout stops looking like a motivation problem. It is Book V, arriving on schedule. We did not only make the deal more fragile. We made the job worse, and then blamed the person for leaving.
The test we skipped
So the argument is not that division of labor is bad. It built real, efficient engines for the businesses it fits, and it deserves the credit. The argument is that it has one precondition, decomposability, and that our industry adopted the factory without ever asking whether the work in front of it was a pin or a relationship. Specialization improves a decomposable task and destroys an indecomposable one. That is the whole test, and almost no one runs it before reorganizing the floor.
Two instincts I have carried for years fall straight out of this, and I finally know why I hold them. The first is that I would rather build a team that closes the whole deal, retail-style, end to end, than run one more assembly line, because keeping the deal in one head keeps the indecomposable thing intact. The second is that when I hire, I would rather recruit brokers than athletes. A commercial real estate broker already runs the entire cycle, origination through close, and already lives in the gaps between parties who do not yet trust each other. That is the whole job, undivided, and you cannot teach it on a conveyor belt, because the conveyor belt is precisely what removed it.
You can put a pin on an assembly line and lose nothing. You cannot put a craft on one without losing the thing that made it a craft. The question was never whether to specialize. It was whether the work in front of you is the kind that survives being cut apart, and complex selling, the kind I have done my whole life, is the kind that does not.