In February 1855, Ralph Waldo Emerson wrote a short entry in his journal under the heading "Common Fame." A man who could grow better corn, he observed, or make better chairs or knives, crucibles or church organs, than anybody else would find "a broad hard-beaten road to his house, though it be in the woods." Thirty-four years later a woman named Sarah Yule published a version she said she had copied from one of his lectures, and somewhere in the retelling the chairs and organs became a mousetrap. Build a better one, the line now runs, and the world will beat a path to your door.
It is the most quoted sentence in American commerce, and it has never once been true on its own.
Emerson knew this, or at least his sentence did. The road in his journal is beaten by people who already know about the corn, and the title tells you how they came to know: common fame, the slow traffic of reputation from one neighbor to the next. The clause most people drop, "though it be in the woods," is not a promise that the woods do not matter. It is an admission that someone has to walk through them first, and that the road exists only because enough people did.
The Accounts of the Better Mousetrap
The sentence is back in fashion, and for a reason that sounds plausible. Making things has become cheap. Software writes software, models draft the campaign, and an answer engine is supposed to read the whole market and recommend the best product to whoever asks. If the machines can find quality on their own, the argument goes, the old expense of being found should shrink, and the money can move into the product.
This month the company with perhaps the strongest claim to having built the better mousetrap published its accounts, and they say the opposite. Anthropic's registration statement for its public offering shows revenue of $16.3 billion for the first six months of 2026, against $1.2 billion in the same period a year earlier. It is hard to imagine a product the market has beaten a faster path toward. The filing even says so in its own words: "Much of the diffusion of Claude has been organic."
Then it shows the cost of the road. Sales, marketing, and partnerships expense rose from $120 million in the first half of 2025 to $3.2 billion in the first half of 2026. Revenue grew about fourteen times, as Fortune noted in its reading of the filing. The cost of reaching customers grew about twenty-six times. Measured against revenue, that line went from roughly a tenth to roughly a fifth, which means that as the product got better and the demand got louder, the company chose to spend a larger share of every dollar on the way to its door, not a smaller one.
The composition matters as much as the total. The filing explains that the line consists primarily of fees paid to channel partners, alongside marketing and advertising, including media and public relations. The channel partners are named elsewhere in the document: the three large cloud platforms, each of which is paid a fee to carry the product to customers who already buy there, and consultancies such as Accenture, Deloitte, KPMG and PwC, who place it inside the companies they advise. In the section on risk, the filing adds a sentence every marketing leader could frame: "The promotion of our brand requires us to make substantial expenditures, which we anticipate will increase as our market becomes more competitive."
The frontier, in other words, does not sell itself. It pays tolls, and it buys media, and it expects to pay more of both.
Two Kinds of Road
Read closely, the filing describes the two roads every company travels, whether it names them or not.
The first is the rented road. It is the marketplace where buyers already shop, the partner who already holds the relationship, the platform that already owns the attention. It is fast and it is wide, and it charges by the traveler. The cloud provider takes a fee on the sale; the search engine takes a fee on the click; the consultancy takes its margin on the implementation. A rented road is often the right road, because it reaches buyers on the day they are ready, but its price is set by someone else and it rises whenever the traffic is worth more to the landlord than to you.
The second is the owned road, and it is the one Emerson was describing. It is reputation, the word of customers who tell other customers, the clarity of a position that a buyer can repeat to her colleagues without your help, and the house at the end of the road that is worth arriving at. It is slow to build and it does not appear on an invoice, which is precisely why it compounds. Every satisfied customer widens it a little, and no landlord can raise the rent.
The interesting thing about the filing is that it shows a company doing both at once and refusing to choose. It credits the organic diffusion and then pays to amplify it. The rented roads bring new travelers; the owned road is why they keep coming back once they have arrived. Its reported net revenue retention of roughly 900 percent among business customers is, among other things, a measure of how well the owned road is holding up under the traffic the rented ones deliver.
The Budget That Treats the Road as Overhead
Set those accounts beside the way most companies fund the same work, and the contrast is uncomfortable. The CMO Survey this year found marketing budgets at their lowest share of company spending since 2021, with total marketing spending up just 1.7 percent over the prior year. It also found that when profits fall short and companies reach for cost cuts, marketing is the expense chosen nearly half the time.
That reflex rests on the mousetrap. It assumes that the product is the asset and the road is overhead, a cost to be trimmed when the quarter is tight because a good enough offer will be found anyway. The most successful product launch of the decade has just argued, in a document filed with securities regulators, that the road is part of the asset. The 19 percent figure is not a benchmark for anyone else; it blends channel fees that most companies book elsewhere, and few businesses grow fourteenfold in twelve months. But the direction is the lesson. When the product got better, the spending on distribution rose faster than the revenue, not slower.
There is a reason for that, and it is the same reason the mousetrap myth is newly seductive. When making becomes cheap, products converge. If a rival can match your feature in a quarter, the feature is no longer what the buyer remembers. What she remembers is who reached her first, who explained the problem in words she could carry into a meeting, and who was already trusted by the people she asked. Abundance in the making raises the price of being found, because there is so much more to find.
The Door at the End
Every road, rented or owned, ends at a door, and for most companies that door is the website and the first conversation behind it. This is where the economics of the road are most often wasted. A company will pay a platform for the click, a partner for the introduction, and a publisher for the mention, and then send the arriving buyer to a page that does not answer the question she came with, or a form that promises a reply in two business days.
The road brought her to the house. Nobody answered the door.
The marketing plan for next year should be written as if the road were an asset with a balance sheet of its own. It should know which roads are rented and what each toll costs per arriving customer, and it should know which are owned and whether they are widening or narrowing. It should treat the door as part of the road, because a visitor who leaves unanswered has cost the full price of the journey and returned nothing. And it should resist the comfortable belief that a strong product excuses a weak path to it, because the strongest product in the market has just disclosed that it does not believe that either.
Emerson was right about the road. He was only quiet about who builds it. The world does not beat a path to the better mousetrap. It walks the road someone paid to clear, and it remembers who was standing at the door.
