Most bid/no-bid conversations are held in the language of judgment — relationships, fit, whether the incumbent is vulnerable — and they are usually held after somebody has already started writing. Judgment is the right tool for the parts of the decision that are genuinely uncertain. It is the wrong tool for the part that is arithmetic, and the arithmetic comes first.
The arithmetic is one line. A bid is worth the chance of winning it, times what winning is worth, less what bidding costs. Written out: EV = p × contribution − cost, where contribution is the margin over the life of the contract rather than the contract value, and cost is the fully-loaded cost of the pursuit rather than the proposal-team hours somebody remembered to log.
What the line looks like
Below is that formula evaluated at a $400k contribution, for two pursuits that differ only in what they cost to chase: a $25k response and a $60k one. Neither number is a benchmark — they are the two ends of what the same opportunity can cost depending on whether it needs new past performance, a teaming agreement and an oral presentation, or does not.
- $25k to bid
- $60k to bid
A model, not measured data: expected value is win probability × a $400k contribution, less the bid cost. Substitute your own two numbers and the crossings move; the shape does not.
The numbers behind this chart
| Win probability | $25k to bid | $60k to bid |
|---|---|---|
| 5% | −$5k | −$40k |
| 10% | $15k | −$20k |
| 15% | $35k | $0k |
| 20% | $55k | $20k |
| 25% | $75k | $40k |
| 30% | $95k | $60k |
| 35% | $115k | $80k |
| 40% | $135k | $100k |
| 45% | $155k | $120k |
The lines are parallel, which is the first useful thing the picture says: raising the win probability and lowering the bid cost are not competing strategies, they move the same number in the same units. The second is where each line crosses zero. The cheap pursuit is worth chasing at a win rate that would embarrass anybody who said it out loud; the expensive one needs better than one in seven before it stops being a way of converting cash into practice.
The break-even is a number you already have
Solve the same line for the probability at which expected value is zero and the whole thing collapses to cost ÷ contribution. That is a division a team can do in the room, on numbers it already knows, before anybody has opened the solicitation.
Arithmetic on the same $400k contribution: break-even win rate is bid cost ÷ contribution. The dashed line is a one-in-four win rate, shown as a reference point rather than as a claim about anybody’s.
The numbers behind this chart
| Cost to bid | Break-even win rate | Wins needed per 10 bids |
|---|---|---|
| $10k | 2.5% | 0.25 |
| $20k | 5% | 0.50 |
| $40k | 10% | 1.00 |
| $80k | 20% | 2.00 |
| $160k | 40% | 4.00 |
Two consequences worth sitting with. A pursuit that costs a fifth of what winning is worth needs a one-in-five hit rate just to break even, which is at or above what many teams actually achieve across a portfolio — so a heavy pursuit is not a bigger bet than a light one, it is a different game with a different qualifying score. And the cheapest way to move the number is almost never to write faster. It is to bid less often, on work where the qualification is already done and the past performance already exists.
Losing is not a thing to appeal
The last place a bad bid decision shows up is a protest, and it is worth knowing what that route actually returns before treating it as part of the plan. The Government Accountability Office publishes its own numbers every year.
| Cases filed with GAO, FY2025 | Protests sustained, of those decided on the merits | Effectiveness rate — protester obtained some relief |
|---|---|---|
| 1,688 | 14% | 52% |
Source: GAO Bid Protest Annual Report to Congress for Fiscal Year 2025. The effectiveness rate counts protests where the protester obtained relief either through a sustained decision or the agency’s voluntary corrective action.
Read those together and the picture is neither "protests never work" nor "protests are a strategy". Roughly half of protesters get something, and most of that something is an agency choosing to take corrective action rather than a finding that the award was unlawful — a re-run of a competition you have already lost once, at your own cost, against a buyer who now knows your pricing. It is a remedy for a procurement that went wrong. It is not a remedy for a bid you should not have made.
Which is the argument for doing the division first. The bid/no-bid decision is the only point in the whole pursuit where the cost of being wrong is zero.