Changing direction is treated as a single kind of decision. Its cost varies enormously with timing, because a company accumulates commitments that a change has to unwind.
Early companies have little to unwind
A young company has a small team, few customers, no meaningful contracts and a product that exists mainly as a set of decisions rather than as an asset.
Abandoning that position costs some months and some morale. Almost nothing built has to be dismantled, because almost nothing has been built to last.
This is why early pivots are common, survivable and frequently the reason a company eventually succeeds at all.
Every commitment raises the price of changing
Hires are made for a particular problem, systems are built for a particular workload and customers are signed on promises about a particular direction.
Each of these is useful while the direction holds and becomes an obstacle when it changes, since the skills, the infrastructure and the obligations do not transfer cleanly.
A later pivot therefore involves redundancies, rewritten systems and difficult customer conversations, all of which consume the resources the new direction needs.
Investors and staff were recruited to a story
People join and fund a company on the basis of a specific thesis. Changing it asks them to re-evaluate a decision they already made and defended.
Some will not follow, and the departures tend to include the people most identified with the previous direction, who are frequently the most capable.
The company then attempts a difficult transition with a weakened team and reduced confidence from the parties whose support it most requires.
Late pivots are often really funding events
A company changing direction late usually needs money to do it, and it is asking for that money without the evidence a new direction normally requires.
Investors evaluating the request are weighing a fresh thesis against a track record of not delivering the previous one, which is an unfavourable comparison.
The realistic options narrow to insiders funding the attempt or a substantial reduction in cost to buy the time internally.
Reversibility is worth protecting deliberately
Companies can keep the cost of changing low by delaying commitments that are hard to unwind: bespoke infrastructure, long contracts and highly specialised hires.
That flexibility has a price in efficiency, and it is usually worth paying while the fundamental questions about the market remain open.
Once those questions are answered the calculation reverses, and the commitments that were dangerous become the ones that build a defensible position.
Recognising which phase a company is in is therefore a practical question rather than a philosophical one, and getting it wrong is expensive in both directions.