Most job searches fail in a specific place, and most people cannot name which place. They send fifty applications, hear nothing, and conclude the market is bad or their CV is weak. Sometimes it is. More often, one of five distinct tasks is being done badly while the other four are fine.
An application is not one piece of work. It is five, and they fail differently.
1. The CV has about six seconds
Recruiters do not read CVs. They scan them, decide, and move on. The often-quoted figure is around six seconds for the first pass. Whether it is six or ten is beside the point: the behaviour is a triage, not a reading.
What survives a triage is different from what survives a reading. A dense paragraph explaining the nuance of your role does not survive. A line that says what you did and what happened as a result does.
The practical test: cover everything on your CV except the top third. Does what remains say who you are, what you do, and why you are worth the next twenty seconds? If it takes a full read to work that out, the document is built for a reader you do not have.
Two habits do most of the damage:
- Responsibility language instead of outcome language. "Responsible for the reporting pipeline" describes a job description. "Cut month-end reporting from four days to six hours" describes you. Both can be true of the same work; only one of them is evidence.
- Unquantified claims. Not everything can carry a number, and inventing one is worse than having none. But "improved onboarding" and "cut onboarding drop-off from 40% to 22%" are different claims, and if you have the second, the first is throwing it away.
2. The parse has to survive before anything else matters
Before a person sees your CV, software extracts text from it. That extraction is where formatting goes to die: two-column layouts can interleave, headers and footers are often skipped entirely, and text inside an image simply does not exist.
This is worth separating from a myth that has attached itself to it.
You will read that 75% of CVs are rejected by ATS software before a human sees them. That number comes from a 2012 sales pitch by a company selling resume optimisation, a company that went out of business the following year and never published a methodology. When recruiters are actually asked, the large majority say their systems do not auto-reject on formatting or keywords at all. The real reason most applications go unanswered is far more boring: a popular role gets several hundred applicants in a few days, and a human never gets to most of them.
So the honest version is narrower than the scary version, and more useful:
- Applicant tracking systems rarely reject you outright.
- They do search and rank, and you cannot be found in a search of text that failed to extract.
- A CV whose contact details lived in the header, and were dropped by the parser, is not "rejected". It is unreachable, which is worse and looks identical.
The two-minute check is the whole defence: open your CV, select all, copy, paste into a plain text editor. (What an ATS actually reads goes through what survives that extraction and what does not.) That is roughly what the software sees. If your job titles are tangled with dates or your phone number is missing, fix that before you touch anything else.
Vocabulary matters for the same mundane reason. If the posting says "stakeholder management" and your CV says "worked with clients", a literal search scores you lower, not because the software judged your skill, but because it was matching a string. Using the employer's words for work you genuinely did is accurate writing. Repeating those words twenty times is the thing that actually gets flagged, and it reads as desperate to the human who opens the file.
3. The cover letter is a research artefact, not a formality
Most cover letters are recognisable as templates from the first line, and a hiring manager reading fifty of them recognises them faster than you would like.
What distinguishes the ones that get read is not style. It is evidence that you know something specific about this employer: a product they shipped, a market they just entered, a problem the posting is visibly written around, and that you have connected your own experience to it.
The structure that works is unglamorous:
- Open on the connection between something true about them and something true about you. Never on "I am writing to apply for".
- Name three of your capabilities and tie each to a problem the role actually has.
- Include one result specific enough to be checkable.
- Stop at 250 to 300 words. Length here signals an inability to be brief, which is itself a data point about working with you.
Note that steps 1 and 2 require reading the posting properly, which is the part people skip. A letter that could be sent to any company in the sector is a letter that says you did not do the work.
4. Interview preparation is rehearsal, not revision
Reading about a company is not preparing for an interview. Preparation means having said your answers out loud, in order, before the room.
The pieces worth actually preparing:
- A ninety-second answer to "tell me about yourself" that connects your past, your present and this specific role. This question opens most interviews and is the one most often improvised badly.
- Structured stories for the recurring themes: leadership, conflict, failure, teamwork. Situation, task, action, result. Not because the acronym is magic, but because under pressure people start in the middle and never reach the outcome.
- A weakness that is a real weakness. The transparent non-answer ("I work too hard") is read as either evasion or a lack of self-knowledge. A genuine limitation with evidence you are working on it is a stronger answer than a fake one.
- Questions of your own that show you thought about their problems. Asked at the end, these are frequently what the interviewer remembers.
The failure mode is having the material in your head but never having said it. The first time a sentence leaves your mouth should not be in the interview.
5. The search itself is a process, not a lottery
The most common way to search for work is to scroll, apply, and hope. It feels productive and produces the least.
A search that works has structure to it:
- A target list, built deliberately, with companies chosen for fit rather than encountered by chance.
- A record of what you sent, where, and when, so follow-ups happen on purpose rather than when you remember.
- A fixed, bounded routine. An hour a day sustained for six weeks beats a frantic Saturday followed by nothing.
- A weekly review, where you look at what produced replies and what produced silence, and change something.
The last one is the one people skip, and it is the only one that improves the others. Fifty applications with no review is one attempt repeated fifty times.
Where the failure usually is
If you are getting no replies at all, the problem is upstream: the CV, the parse, or the targeting. If you are getting replies but no interviews, look at the letter and the fit. If you are getting interviews but no offers, it is the rehearsal.
That diagnosis is worth more than any single fix, because it tells you where to spend the effort.
What RezoCV does with this
RezoCV exists because these five tasks are mechanical enough to be helped and tedious enough that people skip them.
It reads the posting, scores your CV against it and tells you where the gaps are; rewrites your experience in the language of that role; drafts the cover letter from what the posting actually asks for; rehearses the interview with you; and tracks what you sent and when.
One rule runs through all of it. It works from the evidence in your CV, and where something is missing it says so rather than inventing something plausible. Tailoring is choosing which true things to lead with and describing them in the employer's vocabulary, and tailoring a CV to a posting is a reordering problem rather than an invention problem. Inventing is claiming a tool you have not used or a result you did not produce, and it is a bad trade, because it survives the document and fails in the room.
You always click Apply yourself. Nothing is sent on your behalf.
The tools are not the difference. Doing all five jobs, instead of one, is the difference.