A crew opens a trench along a terraced street in a Fenland market town on a Tuesday morning. By Friday the ducting is in, the chambers are set, the footway is patched and three hundred homes are passed.
Not one of them is a customer. Nobody in that street has ordered anything. The money is spent, the permit is closed, and the note in the office records three hundred homes passed as though something had been sold.
Something has been created, which is not the same thing: the possibility of a sale. The build cost is almost identical whether nine houses order or ninety.
Homes passed is a cost, and homes connected is a business
A fibre network is years of spending to make something nobody has asked for yet. The civils are most of the cost. Digging, ducting, chambers, reinstatement and the traffic management around them dwarf the electronics at either end, and a contractor who overruns costs you the same whether the street sells or not.
That produces an unusual set of books. The build sits on the balance sheet as an asset, capitalised because it will earn for decades, and the interest on the money that paid for it is entirely real and due every month. A network can report a growing asset base, a growing home count and a deepening hole in the bank at the same time, and all three statements are true.
The revenue, when it comes, is very good indeed. A connected household pays something in the region of £29 a month, gross margin on that runs above seventy pence in the pound because the line is already built, and people move broadband supplier far less often than they move energy supplier. The trouble is entirely on the way in. Building at any real scale starts at £20,000,000 and can run to £150,000,000 before a single town is finished.
Telecoms retail revenue in Britain is a market of the order of tens of billions of pounds a year. The number of firms actually building networks is far smaller and much harder to pin down: the reliable count is Ofcom's register of operators holding code powers rather than any company register, and the figures used here are estimates rather than an official statistic. VAT, for once, is simple. Telecommunications services are standard rated with no relief of any kind.
The code gives you rights over other people's land, and duties when you leave
Britain stopped licensing telecoms operators individually with the Communications Act 2003. Instead there is a general authorisation: you may provide an electronic communications network without asking permission first, and in return you are bound by Ofcom's general conditions covering things like emergency call access, complaints handling and the way you treat customers. Ofcom can enforce those conditions against you.
The powers that matter to a builder come separately. The electronic communications code gives an operator that Ofcom has designated the right to install and keep apparatus on and under land belonging to other people, on terms that can be agreed or, failing agreement, determined. It is a genuinely unusual privilege, and it comes with duties to make good what you disturb.
Then there is the highway itself, which nobody regulates nationally. Every street opening needs the highway authority's permission. You give notice or apply for a permit, you work to the dates you were given, you reinstate the surface to a prescribed standard, and you remain liable for that patch of footway for a guarantee period afterwards. Overrun your permitted dates and the charges start. Fail an inspection and you come back and dig it again at your own cost, which is the bleakest line in a network builder's accounts.
Take-up per home passed decides whether a street pays
Everything reduces to one ratio: how many of the homes you passed actually ordered, set against what it cost you to pass them.
Below roughly a quarter, a street never pays back. At a third it is a good street. The difference between those two outcomes is rarely the engineering. It is whether you got there first, whether anybody sold to the residents while the trench was still open, and whether a larger network decided to build the same street eighteen months later.
That last one is the real risk. A second network in a town is competing for the same households with an identical product, and take-up that would have reached a third arrives at a tenth. Being first to a street is worth more than being better at anything.
Two things improve the ratio honestly. Selling wholesale as well as retail, so other providers put their own customers on your fibre, fills a street faster than selling under your own brand alone. And pre-orders, where households commit before the dig, turn a speculative street into a nearly certain one, at the cost of building much more slowly.
It dies of its own build rate
The end is not dramatic and it is rarely the regulator's doing. Ofcom can fine you and the council can make you resurface a pavement, but neither of them can kill a network.
The build rate can. The diggers run ahead of the sales team because construction is easier to accelerate than persuasion. The capitalised cost of tens of thousands of unconnected homes accumulates on the balance sheet, the interest on it comes out of cash every month, and the funding runs out something like a year and a half before take-up would have covered it. The assets are real, the customers eventually arrive, and the company that built it is not there to see them.
Every other trade sells something it already has. This one spends for years to create the possibility of a sale and then dies in the gap between the two.
Where this shows up in the game
Running a telecoms network in Business Baron is a monthly argument with yourself about streets. Which ones to build next, whether to build ahead of demand or wait for pre-orders that may never come, and how much unconnected plant you are willing to carry while the interest accrues.
The number on the screen is take-up per home passed against your capital cost per home passed, and it is unforgiving in both directions. Ofcom turns up over the general conditions, the highway authority turns up over a permit and a failed reinstatement, and neither is what finishes you. What finishes you is a spreadsheet full of homes passed, a sales rate that never caught up, and an incumbent that came back to the one town you were counting on. The Telecoms Network page has the build costs, the margins and the take-up assumptions written out.
It is one of the specialist trades, and it needs the full game underneath it. It is in that set rather than the main catalogue because it needs a mechanic nothing else uses: cost incurred today, held on the balance sheet, to create capacity that may or may not ever be bought.