The denominator: pregnancies per year of use
The 2% figure is a pregnancy rate, not a per-act probability. It describes how many people using condoms as their method of contraception become pregnant over a year of use, expressed as a percentage of users. Planned Parenthood frames condoms as protection from both pregnancy and STDs (Planned Parenthood), and the effectiveness of condoms, like most forms of contraception, is assessed in two ways. A distinction that matters for reading any failure number correctly (Wikipedia: Condom).
So when you see “2%,” read it as “about 2 pregnancies per 100 people using condoms for a year,” not “2% chance each time.” That denominator. People, over twelve months. Is what makes the number comparable to other contraceptive methods. It is a population-level rate drawn from real-world use, not a guarantee about any single encounter.
It is not a per-act risk
The most common misreading is treating the percentage as a coin-flip applied to each act of sex. It is not. A per-act risk would multiply across every encounter and produce a very different picture. The published figure is a cumulative annual rate for a group of users. The two-way assessment of condom effectiveness exists precisely because a single percentage can be read in more than one frame (Wikipedia: Condom).
Correcting the misreading changes how you plan. If you have sex a few times a year, your personal exposure is not “2% times the number of acts.” The figure tells you about the method’s performance across a year of typical use in a population. Is the right lens for comparing methods and for understanding what “failure rate” is measuring at all.
What “perfect use” assumes
Perfect-use figures assume the condom is used correctly and consistently every time: a new condom for each act, applied before contact, worn throughout, and removed without spillage. Brook’s guidance is explicit that you use a new one each time you have sex, and that condoms work by stopping sperm in semen from reaching the vagina (Brook). Latex condoms are the most common type and help protect against both pregnancy and STDs (Planned Parenthood).
Perfect use also assumes the condom is compatible with what you put on it. Only water-based or silicone lubricant belongs with latex — oil-based products can damage latex condoms (Planned Parenthood). And if you have a latex allergy, the assumption of “the standard condom” breaks: people with latex allergy should use condoms made from another material, such as polyurethane (Wikipedia: Condom). Each of these is a condition that moves you away from the perfect-use number toward typical-use reality.
The STI exclusion: the figure says nothing about infection
The 2% number is about pregnancy only. It carries no information about sexually transmitted infections. Condoms do reduce STI risk by preventing sexual fluids from being transferred between partners (Brook), and latex condoms specifically help protect against both pregnancy and STDs (Planned Parenthood). But a pregnancy failure rate is not an infection rate, and reading it as one is a category error.
That distinction matters because the two outcomes are measured differently and fail differently. If you are weighing infection risk, the relevant guidance is about protection from STIs and knowing how to protect yourself and your partner(s) (CDC). Testing timing is its own question: STIs can take up to 7 weeks after unprotected sex to show up on a test (NHS). None of that is captured by a contraceptive failure percentage.
How the figure is measured
The number comes from use studies — observational data on people using condoms as contraception over time, counting pregnancies that occur. That is why effectiveness is described as assessable in two ways: the method’s performance under ideal conditions and its performance in ordinary use (Wikipedia: Condom). The gap between those two is where most real-world pregnancies sit.
Because the figure is derived from use in populations, it reflects the behavior of the people studied, including imperfect use. It is not a laboratory measurement of latex integrity. That is also why the number is stable enough to quote but not precise enough to promise: it summarizes a distribution of outcomes across many users and many acts, not a fixed property of the product.
What 2% means across a year of use
Translate it into a year: if 100 people use condoms as their method for a year, about 2 become pregnant. That is the practical reading. It is a small number, and it is smaller than typical-use figures for many methods. But it is not zero, and it is not a per-act guarantee. Condoms provide great protection from both pregnancy and STDs (Planned Parenthood), and that protection is strongest when use is consistent and correct.
If a condom splits or slips off, that is the moment the annual rate stops being an abstraction. Brook notes that if this happens, or you are worried, you may need emergency contraception (Brook). That is the concrete action attached to the failure scenario, and it is the point where the number becomes a decision rather than a statistic.
As a practical boundary: use the 2% figure to compare methods and to understand what is being counted. Pregnancies per year of use. And use the STI and testing guidance for the infection question. The pregnancy rate does not answer.