Opportunity Cost

Why does opportunity cost matter to founders?#

Founder time is the scarcest asset, and the one with no price#

A founder-quarter is roughly 650 working hours. It is not fungible, not purchasable, and not recoverable, and it is the input to the four things nobody else can do: closing the first customers of a new segment, recruiting senior people, setting the strategy narrative, and raising money. Any hour spent elsewhere is drawn from that pool. The practical test before accepting any founder-level commitment: which of those four does this displace, and is it worth more?

Capital allocation is a ranking problem, not an approval problem#

Most roadmap processes ask "should we do this?" — a question with a strong yes bias, because almost every proposed feature has some positive expected value. The correct question compares the same resource block across mutually exclusive uses over the same window. Reframing the meeting from approval to ranking is most of the value of this concept.

It clarifies financing trade-offs#

Bootstrapping preserves ownership and control but rations speed. Venture financing buys capacity and speed at the cost of dilution, governance, and a growth obligation. Neither is free; each simply forgoes a different alternative. Naming the forgone alternative is more useful than arguing about which path is "better."

It disciplines customer selection and pricing#

A low-priced custom contract may clear variable cost and still be a bad deal, because it consumes implementation capacity that would otherwise serve a repeatable segment — and produces learning about one account rather than about a market. That is why ideal customer profile work is an economic exercise, not a marketing one.

How do you calculate opportunity cost?#

1. Name the decision and the scarce resource. Not "what should we build?" but "how should this six-person squad spend the next eight weeks?" State the decision date and horizon.
2. Generate feasible alternatives. Include the status quo and explicit delay. Exclude anything that cannot be funded, staffed, and completed in the window — an infeasible option is not an opportunity cost.
3. Estimate incremental consequences for each: cash in and out, probability and timing, capacity consumed or released, learning produced, reversibility, and effect on future options.
4. Put them on a common basis. Same horizon, same risk treatment, same discount rate. For uncertain single-period outcomes: expected value = Σ(outcome × probability) − required future cost. For multi-period flows: present value = cash flow ÷ (1 + r)^t.
5. Rank, and take the runner-up. The opportunity cost of the chosen option is the value of the highest-ranked rejected option.
6. Find the switching point. Vary the key assumption until the ranking flips, and report that number. It is the most decision-useful output and the most robust to bad estimates.
7. Record the counterfactual in writing. Name the project the chosen option displaced. This is what makes later forecast review possible at all.
8. Revisit when the feasible set changes. New capital, a senior hire, a regulatory event, or customer evidence can change what is possible and therefore what is forgone. A correct decision can become wrong without anyone having erred.

A note on discounting: a startup's opportunity cost of capital is far above the Treasury rates in the Key Facts — those are the risk-free floor, not a startup hurdle. For reference, at the 2026 30-year real rate of 2.0%, $1,000,000 of constant-dollar value arriving in ten years is worth $820,348 today. At a rate reflecting startup risk, it would be worth a fraction of that. The point is not the number; it is that far-future strategic value has to be discounted before it can be compared to a near-term alternative.

Worked example: allocating one squad for a quarter#

One product squad, one quarter, two feasible projects:

Comparing on one year of gross profit, same horizon, undiscounted:

enterprise = 0.60 × ($240,000 × 0.80) − $90,000 = $115,200 − $90,000 = $25,200

self-serve = 0.75 × ($150,000 × 0.90) − $55,000 = $101,250 − $55,000 = $46,250

If the company builds enterprise controls, the opportunity cost is the $46,250 it gave up. The net economic position of that choice is:

$25,200 − $46,250 = −$21,050

That does not settle the decision — it prices the burden of proof. Enterprise controls must be worth at least $21,050 more than the model captures, through strategic access, pricing power, or segment entry, before choosing it is defensible.

The switching point is the useful output. Holding everything else constant, enterprise controls tie self-serve at a close probability of 71.0% rather than 60%. Equivalently, holding enterprise at 60%, self-serve would have to fall to a 59.4% success probability to lose. Both thresholds are inside normal forecasting error, which tells you this is a close call dressed up as a $21,050 gap.

Evidence changes the answer. If a customer offers to prepay, lifting the enterprise close probability to 90%:

0.90 × $192,000 − $90,000 = $82,800

The ranking flips, and the opportunity cost of building self-serve becomes $82,800. Opportunity cost is not a fixed property of a project; it moves whenever the feasible set or the evidence moves.

What this model quietly assumes: that the two projects are genuinely mutually exclusive, that ARR arrives on the same schedule, that one year is the right horizon, and that a 60% and a 75% probability are estimated with equal care. Relax the timing assumption — enterprise revenue landing a year later than self-serve, discounted at even 2% real — and enterprise falls from $25,200 to about $22,900 before any risk adjustment. Relax mutual exclusivity, and the answer may be "do a two-week self-serve test first, then decide."

What are the common mistakes?#

• Adding up every rejected option. Only the best feasible mutually exclusive alternative counts. Summing them inflates the cost of every decision until nothing looks worth doing.
• Ignoring non-cash constraints. Founder attention, senior engineering capacity, launch windows, and organisational focus are usually scarcer than budget, and none of them appear in a P&L.
• Comparing mismatched horizons or risk treatments. A one-year contribution figure against a five-year undiscounted strategic forecast is not a comparison; it is a rhetorical device.
• Treating the status quo as free. Delay consumes runway and can forfeit learning, revenue, morale, and market position. "Wait" is an alternative with a cost, not the absence of a choice.
• Presenting a point estimate as the finding. Report the switching point. If the ranking flips inside your forecasting error, say so.

When does opportunity cost break down?#

The next-best alternative is unobservable — permanently. This is the concept's structural limit, not a temporary data problem. You choose one path, so the counterfactual never runs and never generates evidence. Every opportunity-cost number is a forecast of an unrealised world, which means it cannot be validated after the fact and is easily shaped by whoever writes the memo. Consequences: prefer ranges to point estimates; require the same author to forecast both options; and calibrate on the option you did take, since that is the only one that produces data.

Forecasts are political. The alternatives that reach the comparison are chosen by people with stakes in the outcome. An omitted alternative is a silently understated opportunity cost. Ask who generated the option set before scrutinising the arithmetic.

Choices are rarely fully exclusive. Partial allocation, sequencing, and staged tests often capture value from more than one option. When that is possible, the relevant opportunity cost attaches to the marginal resource block — the next engineer-week — not to an all-or-nothing project.

Strategic options resist valuation. A platform capability may enable future products without near-term revenue. Make the enabling pathway and its evidence explicit rather than assigning an arbitrary strategic premium; an unpriced premium is how any losing project gets approved.

Waiting is not automatically wise. Option value is real when information is arriving and the decision is hard to reverse. But competitor moves, customer windows, and team attrition all make delay expensive. Compare "act now," "stage a test," and "wait" as three distinct alternatives with three distinct costs.