7 min read · Last updated August 10, 2026
- In July 2026 the median duration of unemployment was 10.5 weeks and the mean was 24.9 weeks, a gap driven entirely by a long tail of searches that run past six months.
- A quarter of everyone unemployed in July 2026 had been out 27 weeks or longer, which is past the point where most states stop paying benefits.
- Most states cap regular unemployment insurance at 26 weeks, so the tail of the distribution is the part your benefits do not cover.
- Both figures measure searches still in progress, not finished ones, so treat them as a planning range and never as a forecast.
In this article
- Why the average is more than twice the median
- The three checkpoints that tell you which half you are in
- Sizing the runway: the arithmetic
- What these two numbers actually measure
- Week 26 and the benefit cliff
- Frequently asked questions
Marcus was laid off from a logistics coordinator job on July 6 with $4,100 in savings and a household that costs $2,850 a month to run. He built his plan around eleven weeks, because eleven weeks is the number that comes up first when you search how long a job search takes. Eleven weeks is a real number. It is also the wrong number to fund.
Why the average is more than twice the median
The Bureau of Labor Statistics (BLS) publishes both figures every month in Table A-12 of the Employment Situation release. For July 2026, seasonally adjusted, the median duration of unemployment was 10.5 weeks and the average was 24.9 weeks.
When an average sits at more than twice its median, the distribution has a long right tail. Translated: most people who are out of work have been out for a couple of months, and a large minority have been out for well over half a year. The average is pulled up by that minority. The median cannot see them at all, because a median only reports the person standing in the middle of the line.
That is the whole problem with planning to the median. It describes the outcome you would like. It says nothing about the outcome that bankrupts you.
The three checkpoints that tell you which half you are in
The same BLS table breaks the unemployed into duration buckets, and those buckets are more useful than either summary statistic. They give you three dates to watch, and each one rules something out.
Week 5. In July 2026, 28.2% of unemployed people had been out less than five weeks. If you clear week five still searching, the fast exit did not happen, and you are now planning against the remaining 72%.
Week 15. Add the next bucket and 57.8% had been out less than 15 weeks. Passing week 15 puts you in the 42.2% who are at 15 weeks or more. This is the checkpoint most people miss, because it arrives right about when the initial burst of applications has gone quiet and nothing has replaced it.
Week 27. A quarter of the July 2026 unemployed, 25.5%, had been searching 27 weeks or longer. Reaching week 27 is not a personal failure. It is the single most common place to be after the first six months, and it is the place your benefits are least likely to reach.
Sizing the runway: the arithmetic
Marcus spends $2,850 a month. Annualized and divided by 52, that is $658 a week. His state approved him for $520 a week in unemployment insurance for up to 26 weeks. Here is what each planning horizon actually costs him.
| Planning horizon | Plan to 11 weeks (the median) | Plan to 25 weeks (the average) | Plan to 30 weeks (the tail) |
|---|---|---|---|
| Living costs at $658/week | $7,238 | $16,450 | $19,740 |
| Benefits at $520/week | $5,720 | $13,000 | $13,520 (26-week cap) |
| Gap to fund from savings | $1,518 | $3,450 | $6,220 |
| Covered by his $4,100? | Yes, with $2,582 spare | Yes, with $650 spare | No, short by $2,120 |
| Best for | Nobody. It funds the outcome you want, not the one you have to survive. | The default plan. It matches the current average duration. | Anyone over 50, in a narrow specialty, or relocating. |
The 30-week column is the one that matters, and notice why it breaks. It is not that his costs jump. It is that his benefits stop. The 26-week cap freezes the benefit column while the cost column keeps climbing, so the gap nearly doubles between week 25 and week 30.
Marcus is $2,120 short of a 30-week plan. Spread across 30 weeks that is $71 a week, which is a decision about a phone plan, a streaming stack, and a grocery list, not a crisis. Made in week 1 it is trivial. Made in week 25 it is not available, because there is nothing left to cut from.
What these two numbers actually measure
This is the part almost every article about job search length gets wrong, and it changes how much weight the numbers deserve.
BLS measures how long people who are currently unemployed have been unemployed so far. It is a snapshot of searches in progress, not a record of finished ones. Someone counted at week 3 this month may find work next week or may still be counted at week 40. So 24.9 weeks is not the expected length of a completed job search, and neither is 10.5.
What they are good for is scale. They tell you the tail is real, that it is large, and roughly where it sits. That is exactly what a runway plan needs, and it is all a runway plan needs. Use them to pick the size of the buffer, then stop treating them as a prediction about you.
Week 26 and the benefit cliff
The Department of Labor states that benefits can be paid for a maximum of 26 weeks in most states, with additional weeks available only during periods of high unemployment through the Extended Benefits program. Some states pay fewer.
Line that up against the duration data and the mismatch is stark. A quarter of unemployed people in July 2026 were at 27 weeks or beyond, which is the first week after the standard benefit runs out. The design of the system assumes the median. The distribution does not cooperate.
Two moves follow from that, and both belong in week 1 rather than week 20. First, read your determination letter for your own state’s maximum number of weeks and your weekly amount, and build the table above with your real figures. Second, decide now what you will cut at week 12 if nothing has landed, and write it down. A cut you have already chosen is a budget adjustment. The same cut chosen at week 24 is a forced sale.
If you are early in this, the sequencing in our guide to the first 14 days after a layoff covers what to file and in what order. If you are weighing an offer against your remaining runway, the severance runway math is the companion calculation to this one.
Frequently asked questions
Does a 10.5-week median mean half of job searches end within 10.5 weeks?
No. The median describes people who are unemployed right now and counts only the time they have been out so far. Half of them have been searching 10.5 weeks or less at the moment of the survey. Some of those searches will run much longer, so the figure is not a completion rate.
Should I plan for 25 weeks even in a strong labor market?
Plan for 25 weeks unless you have a signed offer. The average moves slowly, and over-preparing costs you nothing you cannot reverse the week you start work. Under-preparing costs you credit card debt at the exact moment your income is zero, which is the worst possible time to borrow.
What happens to my benefits at week 27?
In most states regular unemployment insurance stops at 26 weeks. Extended Benefits add weeks only when a state hits specific high-unemployment triggers, so they are not something to count on in advance. Check your state’s maximum on your determination letter rather than assuming 26.
Which checkpoint should worry me most?
Week 15. Passing week five is common and passing week 27 is loud enough that you will act. Week 15 is quiet: the early applications have gone cold, no deadline has hit, and it is the easiest point at which to keep doing the same thing for another two months without noticing.