What probability of loss is acceptable on a bid?
A 15% chance of losing money sounds fine until you bid ten jobs. Binomial math shows why the acceptable number depends on how many bets you make and how big the loss is.
There is no universal threshold
People often ask for a rule such as "never quote above a 10% chance of loss". No single number is correct, because the probability tells you how often a bid loses money, not how badly. Two inputs belong next to it: how large the loss can get compared with what your firm can absorb, and how many similar bets you make in a year.
One bid versus ten
Suppose each bid carries an independent 15% modeled chance of loss. For a single job that is a 15% shot at a bad outcome. Across ten such jobs the arithmetic changes:
| Outcome over ten jobs | Probability |
|---|---|
| At least one loss-making job | 80.3% |
| At least two loss-making jobs | 45.6% |
| At least three loss-making jobs | 18.0% |
The expected number of losing jobs is 1.5. At 30% per bid the numbers become 97.2%, 85.1% and 61.7%, and the expected number of losses is three. A per-bid probability that seems modest still means losses are a regular feature of the year, so each one has to be small enough to absorb without damage.
Three questions to set your own limit
- How big is the loss in the bad tail? Look at a high cost percentile at your quoted price, such as P95, and convert it to a dollar loss. A 10% chance of a loss you can shrug off is very different from a 10% chance of a loss that threatens payroll.
- How many similar bets will you make, and are they independent? Jobs for the same client, with the same team or on the same technology tend to go wrong together, so real portfolios are lumpier than the independent case above.
- What is the margin worth if things go well? Accepting more loss probability is rational only when the upside, the strategic value of the client, or a hedge such as a change-control clause pays for it.
Turn it into a written policy
A workable policy has two numbers, not one. For example: "Do not bid above a 15% modeled probability of loss, and do not bid where the P95 loss exceeds a stated amount." Record the exceptions and review them after the job. Over time, compare the loss rate you modeled with the loss rate you got, as described in calibrating estimates with completed jobs.
These figures describe your model's assumptions, not the future, and they are not financial advice. To read the loss probability at a specific price, use the fixed-price project calculator, and see the fixed-price death spiral for what one badly underpriced job does to the rest.
Try the calculation or follow the workflow
- Calculator
Margin & markup calculator
Calculate gross margin, markup, profit and the price needed for a target margin. See why a 25% markup is only a 20% margin. Free, no signup.
- Calculator
Discount impact calculator
See how a client discount changes project profit and gross margin. Calculate the maximum discount that preserves your target margin at the entered cost.
- Guide
Freelance project pricing: from capacity to quote
Build a freelance fixed-price quote from billable capacity, delivery cost, a scope boundary and a target margin. Includes a worked example and review checklist.