How long does a job search take in Switzerland?

What the five months actually count

SECO defines the measure in the notes to its monthly report as the number of days between the reference date — the last day of the month — and the date of registration as a jobseeker, on which the person counted as unemployed. Three things are packed into that sentence, and each of them changes the answer.

First, the clock starts at registration. Someone who applies for three months during a notice period and registers on their last working day starts the official count at zero. The statistic is therefore systematically shorter than the search felt, by exactly the stretch in which most people apply hardest.

Second, it counts only days on which the person counted as unemployed. Interim earnings, a fixed-term assignment or an assigned labour-market measure interrupt the count without interrupting the search. The figure measures a status, not an effort.

Third — and this is the part almost every citation misses — it is a stock measure. It counts the people who were still unemployed on the last day of the month, and it counts the time elapsed so far. Anyone who finds work in month three is no longer there on the reference date and never enters the statistic at three months. "Median 5.0 months" therefore does not mean "half of all searches are over within five months". It means: half of the people currently registered as unemployed have been so for less than five months.

Who is missing from the figure

Switzerland runs two counts side by side, and the smaller one is the famous one. In July 2026, 4.8 per cent of the labour force was registered as jobseekers and 3.0 per cent counted as unemployed. Between those two rates sits a large group: 38.7 per cent of all registered jobseekers are not counted as unemployed, and no duration is computed for any of them.

The Confederation publishes monthly what those people are doing. Of the non-unemployed jobseekers, 40.6 per cent are on Zwischenverdienst — interim earnings, a benefit arrangement that tops up pay from work taken below the insured wage — and 38.3 per cent hold a full- or part-time post. The table’s own footnote says that post may be permanent, terminated or fixed-term. A further 7.5 per cent are in a labour-market measure and 13.6 per cent fall under "other": searching abroad, illness, accident, military or civil service, maternity, or not currently placeable.

Added up, 78.9 per cent of non-unemployed jobseekers are working — 30.5 per cent of every registered jobseeker in the country. Close to a third of officially recorded Swiss job-searching happens from inside a job, and not one of those cases carries a duration.

And that is only the registered part. Anyone who has a job, wants to move and therefore never registers with an RAV appears in none of these numbers — not in the rate, not in the duration. In Switzerland that is probably the most common way a job search happens, and it is statistically invisible. If you ask how long a job search takes and get five months, you have been given the answer for a minority of cases.

Half the registered unemployed are foreign nationals

This matters if you are searching in Switzerland on a permit, because it changes who the figure is describing. In July 2026, 50.1 per cent of Switzerland’s registered unemployed were foreign nationals: 31.1 percentage points EU/EFTA citizens and 19.0 points holders of other passports. Registered unemployment in Switzerland is not a domestic phenomenon with a foreign minority in it; it is split down the middle.

Elapsed duration differs between those groups, though less than the discussion around it suggests. The median stands at 5.0 months for Swiss nationals and 5.0 for EU/EFTA citizens, and at 5.7 months for holders of other passports; the means are 7.0, 6.6 and 7.4. That is a real gap and a modest one — well below the spread between age groups.

The twelve-month change points the other way from what you might expect. Registered unemployment among Swiss nationals rose 8.8 per cent on the same month a year earlier and among EU/EFTA citizens 9.1 per cent, while the "other nationalities" group rose 3.5 per cent. Whatever has tightened in the Swiss labour market over the past year has not landed hardest on non-EU passport holders.

One structural point has no number attached because none exists. If you arrive in Switzerland without a Swiss contribution record, you can register with an RAV and use its placement service, but you may have no benefit entitlement — and the "duration of unemployment" statistic is built on the register, not on entitlement. People who move to Switzerland and search before ever holding a Swiss job are counted differently from people who lose one here, and no published duration separates the two.

What got longer, and what did not

Registered unemployment in July 2026 stood 7.8 per cent above the same month a year earlier. That single figure hides the actual movement, because it is spread very unevenly across the duration bands.

Against the same month last year, the one-to-six-month group grew by 0.7 per cent, the seven-to-twelve-month group by 11.7 per cent and the over-one-year group by 29.9 per cent. The inflow into unemployment is essentially unchanged. What changed is the outflow.

Read as shares: a year ago 58.2 per cent of the registered unemployed had been in that position for one to six months, today 54.4 per cent. Over a year: 13.9 per cent then, 16.7 per cent now. The Swiss labour market has not become harder to enter. It has become harder to leave.

That is the most useful correction to the question. For most searchers, "how long does it take" has roughly the same answer it had a year ago. For a growing minority it has got markedly worse — and the mean and the median have moved mainly because of that minority, not because everyone is searching longer.

The clock starts later than the search does

The gap between when a search starts and when the count starts can be read off the entry statistics. Of all new registrations into unemployment in July 2026, 85.8 per cent came directly out of employment. 10.1 per cent were searching for the first time and 4.1 per cent were returning to working life.

Almost every spell of registered unemployment therefore begins with a termination that is already weeks or months old — Swiss notice periods typically run one to three months. That stretch is the most active phase of many application histories, and the official duration does not count it.

A second figure from the same table qualifies the inflow further. Only 36.3 per cent of monthly entries into unemployment are new registrations. The other 63.7 per cent are status changes by people already on the register who had not been counted as unemployed — a fixed-term assignment ends, a measure runs out. The flow into unemployment is mostly movement inside the system rather than entry into it.

For duration that means the figure is not wrong but is a lower bound on something else. If you use it to plan your own search, add the notice period to it, and do not wait for it to start.

What our own index can add: not duration, but senders

Our data holds adverts, not people, so it can say nothing about how long anyone searched. What it can say is how many independent decisions sit inside a results list — and that is the quantity that governs how quickly a search runs out of material.

Across every open Swiss advert we hold, with apprenticeships, internships and trainee posts removed because they belong to a different market, one employer accounts for 5.6 simultaneously open adverts on average, and 46.2 per cent of employers have exactly one. A hundred lines in a list are, on average, about eighteen senders.

The English-language slice behaves differently from the rest, and in the searcher’s favour. Adverts written in English run 3.5 per employer, against 5.1 for German-language adverts, 4.5 for French and 4.4 for Italian. The five largest advertisers hold 11.3 per cent of the English-language adverts, against 14.4 per cent in German, 27.5 in French and 60.9 in Italian. English-language hiring in Switzerland is spread across many employers rather than concentrated in a few — 50.9 per cent of the employers advertising in English have a single open post.

Software and IT, the field most English-language searchers land in, is the most dispersed of all: 1.9 adverts per employer, 62.6 per cent of those employers with exactly one open role, and the five largest holding 14.6 per cent. Compare that with retail across all advert languages, where five employers hold 57.0 per cent of the field, or construction and trades at 42.5 per cent.

The practical consequence cuts both ways. A dispersed field means the number of attempts you make is close to the number of employers you reach, so volume genuinely converts into chances. It also means there is no shortcut: no handful of large employers stands between you and most of the market, and every application is its own approach.

What can actually be shortened

The starting point is the biggest lever and the only one entirely in your hands. Because the count begins at registration and 85.8 per cent of registrations come out of a job, the notice period is the stretch of a search where most is possible and least is at risk: a salary is still arriving, the CV shows a current post, and the time pressure is at its lowest.

The second lever is width. Pace is set less by how many applications you write than by how many employers your search actually reaches. If five advertisers hold half the adverts in your field, widening the occupation or the area does more than sending more applications to the same senders.

The third concerns the in-between states. Interim earnings or a fixed-term assignment move you out of unemployment statistically — 30.5 per cent of registered jobseekers are in that position. For the search itself this is neither good nor bad, but it does mean your elapsed duration stops appearing anywhere, and you are not finished. Keep the search and the records running as before.

And the fourth is the expectation. When the long band grows by 29.9 per cent and the short one by 0.7, a search that runs into the autumn is not a personal failure; it is the most common way the current Swiss labour market makes itself felt. On binding deadlines and proof of applications, only your RAV’s own guidance counts.

What this question cannot answer

Switzerland publishes no statistic on completed job searches. Everything available is the time elapsed so far by people who are still searching. A figure of the form "an application takes X weeks to an offer" does not exist, and any source quoting one has estimated it.

Our own index cannot close the gap either. It shows what is open today, not how long anyone looked. And it does not reach far enough back to say anything reliable about a span the length of a median search — so we deliberately make no claim here about how much of today’s supply will still be open in five months.

The larger gap is the second one: moving job while employed and never registering with an RAV. It is counted nowhere, neither in the rate nor in the duration, even though it accounts for a substantial share of all Swiss job changes. If you are searching from a job, do not read the five months as being about you: they describe people without income, with reporting obligations, and under a different kind of time pressure.

That leaves the honest formulation the data supports. Six months is normal for a registered Swiss search and is not an alarm signal; a year is the point at which the statistics open a category of their own; and for anyone searching from inside a job, none of these numbers says anything at all.