Can I bypass an ATS altogether?
Updated:
In short
Not around the form — but far more often than you would expect, there is a person standing next to it. Not one of the open Swiss adverts we hold gives an e-mail address as the way to apply; every single one leads to a web page. Yet on 42.5% of the 365 application pages we opened on 29 August 2026 there is a reachable e-mail address or phone number. It is there for questions, not for applications: the dossier still goes through the form.
The other reading of the question — outsmart the software with invisible keywords — we measured rather than assumed. Of seven common hiding techniques, four never reach the PDF at all. The three that do land in exactly the same text layer a human copies text out of. A hiding place a program finds is not a hiding place.
If you are reading this from abroad, one number matters more to you than any other on this page: English-language adverts are the most closed of the four. Only 16.1% of their application pages offer any human channel, against 55.7% of German-language ones.
Share of application pages carrying a reachable e-mail address or phone number, by type of page. Own survey, 29 August 2026: 365 pages drawn from the index, at most two per address; placeholders, data-protection addresses and vendor support addresses excluded.
- No Swiss advert in our index gives an e-mail address as the way to apply — the share is 0%; every one leads to a form.
- 42.5% of the 365 application pages opened carry a reachable address or phone number: 48.5% of employer pages, 67.4% of board and agency pages, 22.7% of applicant systems.
- Only 8.5% of pages name an individual with their own address; a shared mailbox is commoner.
- English-language adverts are the most closed: 16.1% with a human channel against 55.7% for German-language ones, and 5% inside applicant systems.
- Of seven ways to hide keywords, four never reach the PDF; the three that do sit in the same text layer any copy function outputs.
- One-point type is not even invisible — in the pixel comparison it remained on the page as a grey line.
There is no back door in the advert itself
The clearest finding of the whole exercise is a zero. Across the open Swiss adverts we hold, the apply button leads to an e-mail address in 0% of cases. Not rarely — never. Every advert leads to a web page with a form on it.
Where exactly it leads can be classified from the shape of the address: 46.5% stay on an address belonging to the employer, 38.8% go to the address of an applicant system, 14.7% to a job board or a staffing agency. The first figure is an upper bound and the second a lower bound, because many larger employers run bought software behind their own address. For the bypass question that changes nothing — both routes end at the same form.
This is where most imported advice asks the wrong question. "Getting past the ATS" sounds like a shortcut through the procedure. In a Swiss advert that shortcut simply does not exist. The only real question is whether somebody is reachable beside it.
How often a human is reachable next to the form
On 29 August 2026 we opened 365 application pages drawn from our index — mixed across employer sites, applicant systems, boards and agencies, across all four advert languages, at most two pages per address — and read whether the page itself carries a reachable e-mail address or a phone number. Placeholders such as "firstname.lastname@", data-protection addresses and the software vendors' own support addresses were not counted.
The result: 42.5% of all pages offer a human channel. Broken down by where the page sits, a very clear pattern appears. On the employer's own address: 48.5%. On a board or a staffing agency: 67.4%, of which 56.5% carry a phone number — agencies want you to call. Inside an applicant system: 22.7%. Face a system and you are roughly twice as likely to be standing at a door with no bell.
A named individual with their own address appears on only 8.5% of pages. The commoner case is a shared mailbox — info@, hr@, jobs@ — on 14.2%. A phone number appears on 27.1%.
The advert language shifts this more than the industry does: German-language adverts 55.7%, Italian 44.2%, French 41.9%, English 16.1%. The gap survives when you hold the page type constant — among employer-hosted pages, German 58.7% against English 36.4% — and it is starkest inside applicant systems, where English-language adverts reached 5%. If you are applying to Switzerland in English, the closed door is your normal case, not the exception.
The two sentences that settle the question
One cantonal advert, running through a Swiss applicant system, prints both halves of the answer directly under one another: first a note that the outgoing post-holder is happy to answer questions on a direct line, and two lines below, "Bitte beachten Sie, dass nur Online Bewerbungen berücksichtigt werden" — please note that only online applications will be considered — followed by a refusal to accept dossiers from recruitment consultants. Read on 29 August 2026.
That is the separation most pages imply without spelling out. The number is an invitation to talk. The route into the procedure is still the form.
The same pattern in other words: a rehabilitation provider names a contact person with function and e-mail address, then closes by saying it looks forward to the application coming through its online tool. A cantonal hospital describes its process starting from the point at which the application has been submitted through its online portal. In none of these is the human an alternative to the system — the human is the step before it.
There are exceptions, and the most interesting one is real. An apprenticeship advert on Coop's own careers site carried three application routes side by side on the same day: WhatsApp, Instagram and the online form. Precisely where applicants may be thirteen years old, the employer that otherwise runs an applicant system lowers the barrier.
Seven hiding places, measured
The other half of the question is not about the route but about the file: can you feed a system keywords the reader never sees? We set the same CV nine times — once with no keyword block, once with a visible one, and seven times with the same block in a different hiding place — printed each version to A4 PDF through a browser, and turned it back into text with the same extraction library our own CV import uses.
Four of the seven never arrive in the PDF: the block positioned off the page, the block inside a zero-height container, the block set to "display: none", and the block set to "visibility: hidden". The resulting files were byte for byte the same size as the version with no keywords at all — there is literally nothing in them. Rely on one of these and you have a CV without the words you were counting on.
Three arrive, and arrive complete: white text on a white background, one-point type, and a block covered by a white box. In all three, extraction returned exactly the same keywords as the visible version.
That leaves the question of whether you can see them. We rendered every page as an image as well and compared it pixel by pixel against the version without the block. White text: zero differing pixels — genuinely invisible. White box on top: two pixels. One-point type, though: more than six hundred differing pixels. It sits on the page as a fine grey line and is noticeable when someone flicks through.
Why it does not help even when it "works"
A PDF does not have a visible text layer and an invisible one. It has one. That is the layer an applicant system extracts, and it is the layer you get by selecting the page with the mouse and copying — which somebody does at some point in every process, at the latest when the dossier is transferred internally. The hiding place hides from a glance, not from the procedure.
On top of that, systems normally show the people reading them the extracted text, not only the file. At that moment a block of keywords appears that the visible page does not contain. That is not an advantage which goes unnoticed; it is a discrepancy between two versions of the same document, and discrepancies are the one thing comparison reliably surfaces.
Nobody can promise you that a CV gets through a filter — we cannot either. What can be stated cleanly is a mechanism: a CV that uses the advert's own vocabulary matches a keyword-based search better than a generic one. That requires no hiding at all, only copying what the advert already says out loud.
The route the federal government itself names
There is a channel that genuinely goes around something — not around the form, but around the advert. The State Secretariat for Economic Affairs (SECO) writes on arbeit.swiss, under "Speculative applications": "According to experts, jobs are frequently filled via speculative applications. These are not made on the basis of a specific job advertisement and are also called blind applications." It is recommended explicitly as an addition, not a replacement.
The same page names three further channels that do not begin with a form: your personal network, staffing services from executive search through private placement agencies to the RAV (the regional employment centres), and actively following up on an application you have already sent.
For readers relocating: this is not the American informational-interview culture. A Swiss speculative application is a full Bewerbungsdossier — a covering letter, a CV of no more than three pages according to SECO, plus Arbeitszeugnisse (the Swiss employer references), diplomas and certificates, all combined into one PDF. It is sent to a named person at a company that has advertised nothing.
Where the direct route is realistic
A speculative application is not equally worthwhile everywhere, and employer size is a usable signal. 46.0% of the employers in our index have exactly one open advert. Among those, 35.7% of adverts lead into an applicant system; among employers with two to five open roles it is 45.4%.
At the top end the figure falls back to 25.1% for employers with more than a hundred open roles — but that is not a counter-example, it is white-labelling. Large employers run their bought system behind their own address. An e-mail to the group still ends up in the same procedure.
In practice: at a small employer with a single open advert, a direct message has a fair chance of reaching the person who also decides. At a large one it reaches the mailbox that points you back to the form.
If you are registered with the RAV
Here a Swiss rule applies that no international guide covers. Article 26 of the Unemployment Insurance Ordinance requires the insured person to make targeted efforts to find work, "in der Regel in Form einer ordentlichen Bewerbung" — as a rule in the form of an ordinary application. A speculative application is an ordinary application; a phone call that does not turn into one generally is not.
The same article governs proof: it must be submitted "for each control period at the latest on the fifth day of the following month", and the competent office checks the efforts monthly. Miss the deadline without an excusable reason and the efforts are no longer taken into account.
Which form is required, and how a speculative application should be entered on it, is handled differently from canton to canton. The ordinance prescribes the deadline and the check, not the channel. Clarify with your own RAV what counts as proof there before spending a month on routes that do not. Note that the ordinance is published in German, French and Italian only — there is no English version to check against.
What follows from this
The honest answer to "can I bypass an ATS altogether?" has three parts. Around the form: no, in not a single advert we hold. Through the software with a trick: no, because four of seven hiding places do nothing at all and the other three sit in the same text layer a human can copy. Around the advert: yes, and the federal government recommends it.
For the dossier that simply means the work happens on the visible page. The words that matter are in the advert — the job title as it is written there, the qualification, the required language level with its CEFR grade, the Pensum as a percentage. They belong visibly in the CV, not in white type underneath it.
And if the application page does carry a number or an address, that is not a loophole, but it is more than 57.5% of pages offer. A short, specific question about the role — before you send, not instead of sending — is the only way past a form that Swiss employers actually keep open.
Two own surveys. First: 365 application pages, opened and read individually on 29 August 2026, drawn from the open Swiss adverts we hold, stratified by type of target address and by advert language, at most two pages per address; a further 26 pages did not answer and are not counted. Second: nine versions of the same CV, printed to A4 PDF through a browser on 29 August 2026, turned back into text with a widely used extraction library and additionally compared as images. The routing and employer-size figures come from the running evaluation of the open adverts we hold, as of 28 August 2026, and are reported as shares; classification follows the shape of the target address, so the "employer page" share is an upper bound and the "applicant system" share a lower bound.
Check how your CV fits a job — and the Swiss market
Free checker: languages, work permit, skills and more — scored against a specific posting.
ATS CV checker: against a specific advert, freeSources
- SECO / arbeit.swiss — Tips on applying for jobs, "Speculative applications" (read 29 August 2026)
- Fedlex — Unemployment Insurance Ordinance (AVIV), Art. 26, version of 1 August 2026 (German text)
- pdf.js — the extraction library used for the test (retrieved 29 August 2026)
- SECO / arbeit.swiss — overview for jobseekers
Related questions
What our job index says about the Swiss market
Computed live from our own index, not quoted from a study. Shares only, as of today.
Language the advert is written in
- Deutsch
- 60%
- English
- 23%
- Français
- 13%
- Italiano
- 3%
Of adverts that state a language requirement, the share asking for
- Deutsch
- 70%
- English
- 43%
- Français
- 21%
- Italiano
- 3%
19% posted in the last 7 days · Largest markets: Zürich 18% · Bern 10% · Genève 5% · Basel 5%