The median duration of unemployment was 10.5 weeks in July 2026, against a 2019 average of 9.2 weeks and a peak of 25.2 weeks in June 2010. The median is the better planning figure because the mean is dragged upward by a long tail of extended searches. Neither describes someone searching from a job they already hold, who is not counted at all.
The number, and which version of it to use
The median duration of unemployment was 10.5 weeks in July 2026, against a 2019 average of 9.2 weeks and 9.4 in February 2020. The series peaked at 25.2 weeks in June 2010.
Two figures get published: a mean and a median. The mean runs well above the median because a minority of very long searches drag it upward, so the median is the one to plan against and the mean is the one that gets quoted for effect.
What the 2010 peak teaches
The recession officially ended in mid-2009. Median unemployment duration peaked a full year later, in June 2010, at more than twice today’s level.
That lag is the most important thing this series has to teach. Long-term unemployment rises late and falls slowly, so “the downturn is over” and “searches have got shorter” are separated by a considerable gap. If you are searching during a recovery, the conditions you face are the previous year’s, not this quarter’s.
Who is not counted
Anybody searching while employed — a large share of all job seekers, and usually the ones with the shortest searches, because they can wait for the right role and they interview from a position of not needing to.
New entrants and people returning after time out are also counted differently from those who lost a job. So this figure describes unemployed searchers specifically, and the population it describes is the one having the hardest time by construction.
That cuts both ways for you. If you are employed and looking, your realistic timeline is probably shorter than 10.5 weeks. If you are not, it is a fair benchmark.
What actually lengthens a search
Seniority. There are fewer senior roles and their processes are longer, with more interview rounds and more people who have to agree. A director-level search taking three times as long as an analyst-level one is normal rather than a sign of trouble.
Narrow specialization. A field with few openings anywhere means waiting for one rather than choosing among several, and no amount of effort creates a vacancy.
A constrained local market. If your occupation is thin where you live, the search is really a decision about relocating disguised as a job hunt.
Changing field. Every application carries a persuasion step that an in-field candidate does not need, and that step fails most of the time even when it is done well.
Planning against the distribution, not the average
Use the median for your expectation and the long-tail share for your finances. A budget built on the mean is built on a number distorted by exactly the outcome you are trying to avoid, and a budget built on the median leaves you exposed if you land in the tail.
The practical version: plan your time around roughly three months and your money around roughly six. That gap between the two is not pessimism, it is what the shape of the distribution actually requires.
The 27-week line
The share of unemployed people out of work for 27 weeks or more is published alongside the duration figures, and it behaves differently from the median. It is the measure that matters most, because long-term unemployment is where the real damage happens — skills atrophy, networks thin, and employers start treating the gap itself as a signal.
If that share is rising while the median holds steady, the market is bifurcating: most people are fine and a growing minority are stuck. That is a different situation from a general slowdown and calls for a different response, mainly urgency.
What this figure cannot tell you
Anything about your occupation or your metro, which is where the variation lives. A national median across every occupation in the country is a shape rather than a prediction.
It also says nothing about the quality of the job at the end. A short search that ends in a step down and a long one that ends in a step up look identical in this data, and they are not the same outcome.
Common questions
How long does a job search take?
The median unemployment spell was 10.5 weeks in July 2026, against a 2019 average of 9.2 weeks.
How bad is that historically?
Longer than before the pandemic and far from severe. The series peaked at 25.2 weeks in June 2010.
Why use the median rather than the mean?
The mean is dragged up by a minority of very long searches. The median is where half of spells are shorter.
Does it include people searching while employed?
No, and they are a large share of job seekers, usually with the shortest searches.
How should I budget?
Median for expectation, long-tail share for finances. Planning on 2019's 9.2 weeks is optimistic for 2026.
How long does a job search take?
The median unemployment spell was 10.5 weeks in July 2026, against 9.2 in 2019 and a peak of 25.2 weeks in June 2010.
Should I use the mean or the median?
The median. The mean is dragged up by a minority of very long searches, which makes it a poor planning number and a good headline.
How should I budget?
Plan your time around roughly three months and your money around roughly six. The gap reflects the long tail in the distribution.