In the early months of 2020, as gyms across the United States closed under pandemic restrictions, Peloton became one of the fastest-growing consumer companies in the country. Demand for its stationary bikes and subscription classes outpaced the company's manufacturing capacity for months at a stretch. Winning that attention was a single event. Keeping it turned out to be a recurring cost, paid in instructor development, community moderation, and structured retention programs, and Peloton had built out only the first of the two disciplines in time.
As pandemic restrictions eased and gyms reopened, the company found itself carrying manufacturing capacity, inventory, and headcount built for a demand spike that had already passed. Subscriptions were cancelled at scale. Inventory piled up in warehouses. Leadership changed more than once in the years that followed, including the departure of co-founder and CEO John Foley from the chief executive role in February 2022. Peloton had not lost the attention it won in 2020 because people had stopped caring about home fitness. It lost it because success itself had created an obligation nobody had funded, and the obligation came due anyway.
Every Launch Creates Maintenance
Attention decays at a measurable rate unless deliberate effort continually reinvests in retaining it, the way an untouched sum of money quietly loses real value to inflation even without anyone withdrawing from the account. This is not a claim specific to marketing. It is closer to entropy applied to anything built by people: a marriage, a friendship, a government, a media brand, a piece of software, all decay by default unless someone is actively paying the cost of keeping them from decaying. Success does not exempt anything from this. It just adds more of it that now needs maintaining.
Peloton's decline, in this framing, was less a sudden collapse than an unpaid tax accumulating quietly across a year in which the company was still publicly celebrating the demand numbers that had won its original attention.
Ray Kroc's Operations Manual
Ray Kroc joined the McDonald brothers' single restaurant operation as a franchise agent in 1955 and bought out Richard and Maurice McDonald's ownership stake in 1961. What made McDonald's capable of scaling into a global chain was not its advertising. It was an operations manual, expanded over the following years into hundreds of pages of specific instruction, dictating exact cooking times for french fries and exact standards for how a franchise location should be run, so that a McDonald's in one city behaved identically to a location a thousand miles away. The company opened Hamburger University, a dedicated training facility for franchise operators, in Elk Grove Village, Illinois, in 1961.
Toyota built a version of the same discipline into an entire philosophy of production. Beginning in the years after the Second World War, Toyota engineer Taiichi Ohno developed what became known as the Toyota Production System, built around continuous, deliberate small improvements, kaizen, and a manufacturing process designed to expose and correct problems immediately rather than let them accumulate. A 1990 MIT study of the system, published as The Machine That Changed the World, formalized it globally as lean production, and it remains one of the most studied management systems in the world specifically because it treats maintenance and improvement as the actual product, not overhead sitting beside it.
Kroc's operations manual and Ohno's production system are the same law wearing two different industries: a strong product or a temporary spike in demand only produces a durable business when a disciplined system exists underneath it, built to keep delivering the same experience the millionth time as the first, and rebuilt continuously rather than assumed to hold on its own.
Growth Creates Recurring Obligations
Every customer a company wins creates an ongoing obligation to support them. Every employee hired creates an ongoing obligation to manage them. Every audience built creates an ongoing set of expectations that has to keep being met, at the same standard, indefinitely. None of this is optional overhead attached to growth. It is what growth actually is, once the first spike of attention has passed and the accumulated obligations start needing to be paid down every single day rather than celebrated once.
The companies that survive are rarely the ones that launch best. They are the ones that maintain best, long after launching has stopped being interesting to anyone outside the company.
What Machines Might Change
As distribution and creative production get cheaper through AI tools, launching well is becoming easier for nearly everyone at the same time, which means it stops functioning as a real differentiator. What remains scarce, and what should increasingly become the actual advantage, is the discipline to keep paying the maintenance cost long after a launch has stopped generating its own attention. The clearest way this forecast fails is if AI also resolves a meaningful share of that maintenance work itself, automating retention and community management cheaply enough that the obligation stops being expensive to carry. If that happens, the underlying law will not have changed. The scarce resource will simply have relocated to whoever builds the systems doing that maintenance most reliably.
Engine
Growth is the easy part of scaling. Every launch creates maintenance the launch itself never had to pay for, and the bill comes due regardless of whether anyone budgeted for it.
The first implication is funding the maintenance layer at the same time as the launch, not after it. Peloton's manufacturing and marketing scaled together in 2020. Its retention infrastructure didn't, and that gap is where the company's momentum leaked out.
The second is measuring decay, not just peaks. A launch week's numbers say nothing about whether the gain is being retained or quietly taxed away. The more useful number is whether this month's baseline is holding against last month's, long after the initial excitement has worn off.
The third is treating maintenance as a real discipline with its own budget and its own hires, not a byproduct of a good marketing team.
Ray Kroc's operations manual and Taiichi Ohno's production system were never marketing decisions.
They were the actual mechanism that turned one good moment into a business still standing decades later, long after the moment that first won it any attention had been forgotten by everyone except the people still paying to maintain it.
