The Escalation Continuum: Why Sunk Costs Matter More Than Heuristics

The sunk cost fallacy is not a cognitive error—it is a rational response to how we structure commitments.

This distinction matters because it reframes what behavioural scientists have spent decades treating as a bias. Kahneman and Tversky's work on prospect theory gave us the language to describe why people throw good money after bad. We learned to call it irrational. We built interventions around correcting it. But the real mechanism operating beneath sunk cost behaviour is far more interesting than a simple failure of logic.

Consider what actually happens when someone continues funding a failing project. The decision-maker is not ignoring the sunk cost—they are responding to the structure of escalation that the sunk cost creates. Each incremental investment feels smaller than the total already committed. A £50,000 loss feels catastrophic; a £5,000 additional spend to "save" that investment feels manageable. The psychology is not about irrationality. It is about how costs are framed across time.

This is where the reference documents on decision physics become crucial. The AGE framework—Anchoring, Gradation, and Escalation—describes how decisions accumulate rather than stand alone. Sunk costs work precisely because they anchor future decisions. Once you have committed £50,000, the next £5,000 is not evaluated against the original opportunity cost. It is evaluated against the loss you will crystallise if you stop. The gradient matters more than the absolute value.

The mistake in most behavioural interventions is treating sunk costs as a heuristic problem. Researchers have tried to "debias" people by making them aware of the fallacy. Workshops teach executives to ignore sunk costs. But awareness alone does not change the structure of how costs are presented. If you break a £50,000 commitment into ten £5,000 tranches, each decision point feels like a small, defensible choice. The total commitment remains irrational, but each individual step appears rational within its local context.

This is why the escalation continuum matters. It is not a single decision point where someone makes a bad choice. It is a series of micro-decisions, each one individually justifiable, that collectively produce the outcome we label as "sunk cost fallacy." The person is not failing at logic. They are responding predictably to how the decision environment is constructed.

The implications for strategy and marketing are substantial. If you want to understand why customers remain locked into failing products, why teams defend underperforming initiatives, or why organisations continue investing in legacy systems, you must look at how costs are distributed rather than how they are totalled. A subscription model that breaks annual costs into monthly payments will retain customers longer than a single annual charge, not because customers are irrational, but because the escalation structure makes each renewal feel like a smaller, more defensible decision.

Similarly, in organisational contexts, projects that require staged funding will accumulate more total investment than those requiring upfront capital, even when the expected returns are identical. The escalation continuum creates psychological momentum that pure cost-benefit analysis cannot explain.

The deeper insight is that sunk costs are not a bug in human decision-making—they are a feature of how we process commitment over time. We evolved to make decisions incrementally, with each choice building on previous investments of effort, resources, and reputation. In stable environments, this works well. In volatile ones, it produces the escalation patterns we observe.

The practical question is not how to eliminate sunk cost thinking. It is how to restructure decision environments so that escalation patterns align with strategic intent. If you want people to abandon failing projects, do not simply tell them to ignore sunk costs. Instead, reset the decision frame. Create natural breakpoints. Require explicit recommitment rather than passive continuation. Change the gradient of the escalation continuum.

This is the real work of decision design—not correcting individual biases, but understanding how the structure of choice produces predictable outcomes.