A good application takes time. People who use Prizm say it takes them thirty minutes to an hour when they tailor their CV and cover letter themselves. Across dozens of job ads, that adds up to a week of work. So the question that matters comes before the CV: is this job ad actually worth that hour?
Since August 2026, Prizm has analysed more than 1,700 real job ads, each one read against the background of the person who wanted to apply. Here is what those analyses show, and how you can use it, with or without us.
One job ad in three isn't worth applying for
For every job ad, Prizm gives one of three recommendations: go for it, try it eyes open, or pass on this one.
- Pass on this one: 32% of job ads. The profile and the role are too far apart for an application to stand a reasonable chance.
- Try it, eyes open: 55%. There is a real foundation, along with gaps you will need to own.
- Go for it: 13%. The profile matches what the role asks for.
In other words, nearly one in three job ads people were about to apply for wasn't worth it. And barely one in ten is an obvious fit.
This is not a judgement on anyone. People chose these ads themselves, usually in good faith, because the job title looked like their line of work. But a title says very little. The list of requirements, the expected level and the context are what decide.
Four job ads in ten hide a dealbreaker
An analysis finds five gaps on average between someone's background and the job ad. Most are significant but fixable: a skill to bring forward, the ad's vocabulary to reuse, an experience to present in a different light.
Some gaps, however, are dealbreakers: a degree required by regulation, years of experience far above what the person has, an incompatible type of contract. 39% of the job ads analysed had at least one.
A dealbreaker can't be fixed with a better cover letter. It is visible at first glance, to a recruiter as much as to screening software. It is the first thing to look for, before anything else.
A third of applications fall at the first hurdle
For every job ad, the analysis also estimates the chance of getting through the first screening, the one that decides whether an application gets read in full.
- Low chance: 33% of job ads.
- Moderate chance: 50%.
- Good chance: 17%.
First screening is harsh because it is fast. A few seconds, or software looking for the ad's own words in your CV. An application that doesn't show right away that it covers the essentials goes no further, however good the rest may be.
Knowing when to walk away saves time
The figure that taught us the most is about what people do with the recommendation.
When the analysis says go for it, two people in three go on to prepare their application. When it says pass on this one, seven in ten stop there.
Far from being discouraging, walking away is a real time saver in a job search. Every job ad set aside early is an hour that doesn't go into an application with no future, and goes instead to a role where the application has a chance of being read.
That is the whole point: not sending more applications, but sending better ones, where they count.
How to decide for yourself in five minutes
You can do this sorting without any tool. Before you open your CV, read the job ad again with these five questions in mind.
- Is there a non-negotiable requirement I don't meet? A regulated qualification, a security clearance, a driving licence, a language used every day. If so, and you don't have it, pass on this one.
- Is the experience they ask for within reach? A gap of a year or two can be made up for with results. A five-year gap, or a much higher level of responsibility, rarely can.
- Does the contract suit me? Apprenticeship, internship, permanent, part-time, location, travel: a mismatch here often rules you out before your CV is even read.
- Can I find at least the first three responsibilities of the role in my own background? They are usually the ones that matter most to the person hiring.
- Do I already use the ad's words to describe what I do? If you do the same thing under a different name, that can be fixed: use their words, without inventing anything.
If your answers to the first three questions are good, the application is worth its hour. If one of them is clearly bad, save your energy for the next one.
What these figures don't tell you
These analyses measure how well a background matches a job ad, as far as it can be read from the documents. They don't say who got hired: an interview also depends on the number of candidates, networks, timing and the person across the table. They cover job ads that people chose themselves, not a representative sample of the job market. And a "low" chance is not zero: a recommendation remains a recommendation.
Method
Analyses carried out by Prizm between 12 August and 23 September 2026, on more than 1,700 real job ads, each read against the CV of the person who wanted to apply. Test accounts and team accounts are excluded. Figures are proportions calculated across the whole set: no job ad, company or person can be identified.