Rising Up recognized by Deeptech...
Cost of a bad hire in 2026: €30,000 APEC, often over €60,000 in reality. Why the number is underestimated and 4 scientific levers to reduce the risk.
You just learned that one of your recent hires will not pass the probation period. You know it costs you money. But how much, exactly?
The official figure has been circulating for years: several tens of thousands of euros. Most HR directors quote a similar figure in their executive committees. Very few realize it drastically underestimates the real cost.
Why? Because this amount is only the tip of the iceberg. The rest is what happens next: the manager who disengages, the team that compensates, the cascade of departures, the bitter Glassdoor review, the next hiring cycle starting with an already burned-out recruiter.
In this article, we will break down together:
And at the end, we will see where to start concretely, without changing your entire recruitment method overnight.
Before looking at the precise numbers, let's set the scene.
According to Apec's 2026 survey on executive recruitment practices, 41% of companies hiring executives now formally assess soft skills, up 9 points compared to 2022. Professional development budgets are increasingly oriented toward behavioural competencies. The stakes have more than doubled in less than a decade.
Yet, these are precisely the skills that are least well assessed in interviews. The result? A probation failure rate in France that structurally oscillates between 15 and 20% depending on the sector (Pôle emploi, executive sector studies). Out of 10 executive hires, on average 2 will not pass probation.
Now let's do the math: if your company hires 50 executives per year, that means 10 statistical casting errors every year. At several tens of thousands of euros each on the low estimate, you already reach hundreds of thousands of euros of visible annual cost just for bad hiring.
And that's only the portion that shows up in the P&L. The real cost, as we will see, is much higher.
The classic estimate adds up three blocks of directly traceable costs. Let's take them one by one, with concrete examples on an executive position paid €55,000 gross annually.
Fully loaded salary during the non-productive probation period (before the person generates their full value), recruitment fees (agency, jobboard, sometimes LinkedIn Premium Recruiter), onboarding, initial training, equipment, SaaS tools access rights.
👉🏼 Over 6 months, you are already at €12,000 to €18,000. And that's without counting the prorated bonus, travel expenses, and the possible cost of a recruitment agency (which can represent 15 to 25% of the annual salary alone).
If you went through an agency for a €55,000 position, count €8,250 to €13,750 more just for the agency service. Whether the person stays 3 months or 3 years, you paid for it.
This is the block that HR directors most often forget in their internal calculations.
The manager's time spent onboarding, coaching, shadowing, debriefing a candidate who will not stay. The colleagues' time spent compensating, redoing, doubling up, or bearing the poor fit. Team meetings slowed down by the need to make explicit what should be implicit.
👉🏼 At €600 per management day on average in France (a conservative figure for a manager at €80,000 gross fully loaded), 30 scattered management days over 6 months represent a hidden cost of €18,000.
And these 30 days are a low estimate. On a poorly filled operational position within a team of 6 people, the negative halo effect on collective productivity can represent 5 to 10% of the team's total time during the period. Multiply by 6 collaborators: you easily reach 60 to 100 diffused team days.
This is the most volatile block. The position left vacant after the departure. The time to restart a process (on average several weeks for an executive position in France). The productivity that the team does not generate during this time. The client project that slips, the deal that is not signed, the product roadmap that drifts.
It is also the block most often forgotten in HR calculations, because it is invisible: it never appears as an expense. It appears as a shortfall. And shortfalls are almost never accounted for.
On a Business Developer position whose annual objective is to generate €500,000 of pipeline, a month and a half of vacancy represents around €60,000 of pipeline not generated. On a Lead Tech, that is a feature that ships 1.5 months later, and an impact on product retention that translates into thousands.
The classic estimate adds up what is directly measurable. Three additional factors push the bill up, and they are almost never in HR estimates.
A bad hire is not just one person leaving. It is a team that becomes destabilized.
Colleagues close to a poorly adjusted new team member show an increased risk of disengagement within the following 6 to 12 months. A newcomer who does not align with the culture or the requirements of the position generates a friction signal that spreads across the team: several close collaborators may drift into disengagement, raising the risk of them leaving in the following months.
👉🏼 Conclusion: the secondary replacement cost can easily double the initial bill when this ripple effect is included.
Glassdoor, Indeed, LinkedIn reviews durably crystallize the failed candidate experience. A bad departure often generates a sharp review. A sharp review lowers your employer Net Promoter Score. A declining employer NPS mechanically increases the acquisition cost of your future applications.
And it is cumulative. Over 3 years, an eroding employer brand can increase candidate acquisition cost by 20 to 40% according to LinkedIn benchmarks. It is a diffuse effect, never charged to a specific recruitment, but very real in HR budgets.
An effect rarely integrated in HR calculations. And yet: if you had to quantify it, you would probably put it at €3,000 to €8,000 per bad departure.
After a failure, recruiters tend to compensate. Either by excessive caution: lengthening decision times, losing good candidates through slowness, multiplying interview rounds for reassurance. Or by inverse excessive confidence: abusive shortening of the process to close the vacant position that weighs, return to pure instinct, gut-based hiring.
Both degrade subsequent hires. And no one sees it, because the degradation is smoothed over time.
It is an organizational opportunity cost measured in the average quality of future hires, never as a line in Excel.
If this order of magnitude has been public for years, why do HR directors continue to massively minimize the cost of a bad hire?
Three documented cognitive biases explain this persistence. You will probably recognize them.
"This time will be different."
Every new process restarts with an implicit belief that the past mistake will not repeat itself. Yet, nothing in the method has fundamentally changed: same screening criteria, same interview grid, same end-of-process intuition, same biases.
This is what Daniel Kahneman calls the "illusion of validity": we believe our predictions about humans are reliable, when empirical studies show that unstructured interviews are barely more predictive than random selection when it comes to assessing soft skills.
Prestigious degree. Experience at a well-known brand. Polished LinkedIn profile, with pro photo and a well-crafted tagline. These signals activate a decision shortcut that overestimates the probability of behavioral success.
Yet, and this is the key point: behavioural skills predict professional success, contribute to it causally, and can be developed (research by James Heckman, Nobel laureate in economics). Not on the hard skills visible on the CV. Not on the school name. Not on the previous employer brand.
A candidate from a Tier 1 school can be a poor communicator, cruelly lack emotional regulation, or have very low mental flexibility. The CV will never tell you. The classic interview either, because the candidate knows the right answers to classic questions.
When a position is vacant, the daily opportunity cost creates urgency. And this urgency pushes to accept the "not ideal but okay" candidate, the one who "does the job without us really being convinced".
This pressure is rarely quantified as a bad hire risk in committees. Yet, it is one of the first predictors of bad casting according to feedback from our clients (Google, Adobe, Orange).
When an HR director tells us "we hired too quickly", we generally hear: "we did not run our objective measurement process because we didn't have time, and intuition spoke." And intuition is biased 80% of the time.
How many companies conduct, 6 months after each hire, a structured debrief on the quality of the process that led to this hire?
Very few. HR does the debrief if the recruitment fails (and even then, often orally, without data). But if the hire "passes", no one looks at whether the decision could have been better, or whether a better candidate was missed. Organizational learning on recruitment is almost non-existent.
Without data, no learning. Without learning, the error cost remains constant year after year.
Hiring remains a human act of decision-making. Nobody proposes to replace it with an algorithm. But it can be structured to reduce the variance of results.
Here are the 4 scientific levers that Rising Up applies with its clients Google, Adobe and Orange.
It all starts before the interview. It all starts even before the job posting. It all starts with the job description.
A classic job description lists the expected technical skills: 5 years of Python experience, B2B SaaS mastery, fluent English. It almost always omits the implicit soft skills that make a decisive part of real performance: ability to work in ambiguity, tolerance for pressure, quality of written communication, emotional regulation in internal conflicts.
Rising Up uses CoreSkills AI, our proprietary AI, to scan each job description and reveal the expected soft skills, explicit and implicit, aligned with a scientific reference framework of 18 soft skills across 3 families (Leadership & Collaboration, Innovation & Communication, Operational Efficiency).
It is our AI that performs this scan, not a third-party tool: it has been trained on the Rising Up scientific framework, calibrated by the scientific team of Dr Nawal Abboub. Concretely, you get a complete semantic mapping of your job description, an alignment score between what you explicitly ask and what really drives success on the position, and rewording suggestions to attract the right profiles while avoiding biased self-selection.
This is exactly what we deployed at Adobe on the "Communicate From Everywhere" program: before measuring candidates' soft skills, we first mapped the managerial job descriptions with CoreSkills AI to make explicit what "communicating remotely" really meant within their culture.
Classic psychometric tests (PAPI, 16PF, Hogan, MBTI, DISC) rely on declarative self-assessment. The candidate ticks what they think they are. The result? Everyone overestimates themselves in interviews. The best even make an art of it.
This is called the social desirability bias: candidates answer what they think the employer wants to hear, not what they actually do. On the 18 key soft skills in business, the gap between declaration and real behavior oscillates between 20 and 40% according to psychometric meta-analyses.
Rising Up's Soft Skill Scan changes the paradigm. Developed by a company recognised as Deeptech by the European Innovation Council (EIC) in 2023, the Scan combines two complementary dimensions:
Not what the candidate says they do. What they actually do, facing situations whose right answers cannot be guessed in advance.
The candidate completes the Scan in 20 minutes, from any device, without possible preparation. You get a behavioral profile on the 18 key soft skills, presented as strengths and points to watch, directly usable as an interview preparation grid.
Rather than approaching the interview with a partial intuition, the manager arrives with a more objective viewpoint on the candidate, structured on a clear scale.
The recruitment test THE SCAN produces this result by crossing the soft skills actually expected on the position (output of Lever 1) with the candidate's behavioral profile (output of Lever 2). The result is expressed on a simple three-level scale:
The interview then becomes what it should always have been: a conversation focused on the "To explore" areas, not a discovery of the candidate from scratch. The manager saves 30 to 45 minutes per interview and arrives at the final decision with a qualitatively superior reading.
It is a decision aid, provided to the manager to inform their choice. The tool prepares, enriches and orients the interview: it does not replace it. The final decision always belongs to the manager, after analysis of the elements provided by the Scan and conduct of the interview.
Here is the most chronic blind spot in HR: inter-recruiter variance.
When 3 recruiters evaluate the same candidate according to 3 different implicit grids, the result is partly random. The same candidate will come out "Compatible" with Recruiter A, "Uncertain" with Recruiter B, "Incompatible" with Recruiter C. This variance pollutes all decisions in your pipeline.
Aligning the entire HR team on the 18 Rising Up soft skills, and calibrating evaluation anchors on precise observable behaviors, drastically reduces inter-recruiter variance. Everyone speaks the same language. Decisions become, comparable, defensible in executive committee.
This is typically what we accompanied at Orange Cloud for Business on their managerial team: before alignment, several managers evaluated candidates with different implicit grids. After calibration on the Rising Up reference framework, they use the same evaluation language, hiring decisions become faster and defensible in committee.
With an average hiring rate of 15% per year (15 new positions), and a probation failure rate of 15-20%, count 2 to 3 casting mistakes per year. At €30,000 to €60,000 each on the real cost, you are structurally between €60,000 and €180,000 of annual error cost on hiring alone, excluding HR time.
No. The Soft Skill Scan and CoreSkills AI integrate upstream of the interview, without replacing your ATS or your sourcing channels. You keep your current process; we enrich it with an objective layer of behavioral measurement, without breaking what works.
Over the 8-week pilot, you measure the impact directly on the 5 targeted hires (better prepared interviews, objective scoring, team calibration). The long-term impact on the probation failure rate is measured over 6 to 12 months, in coherence with classic HR cycles.
The Scan is a decision aid for the manager, not an oracle. It reduces variance, it does not eliminate it. The final decision remains that of the manager, after analysis of the Scan and conduct of the interview. No scientific tool can predict success in position at 100%, because success also depends on context (manager, team, market situation). But it shifts your process from an average 20% failure zone to a lower zone.
For a company hiring 5 executive positions over 8 weeks, the statistical cost of a casting mistake often exceeds €60,000 (a bad hire added to its cascade turnover).
This is exactly the window targeted by the Rising Up pilot program:
🎯 150 behavioral evaluations over 8 weeks, on the candidates of your 5 targeted positions
🎯 The CoreSkills AI analysis of your 5 job descriptions (acquired for life, even if you stop mid-course)
🎯 €3,000 excl. tax paid in one go at signature, valid until September 30, 2026 (then €4,500 excl. tax)
🎯 Limited to 5 companies per month
The asymmetry is clear: a single avoided bad hire covers 10 times the cost of the pilot. And even if you stop mid-course, the mapping of your 5 job descriptions remains an asset that structures all your future hires.
Send us a message 📩 at hello@risinguparis.com and we will organize a 15-minute demonstration for you within the week.
From recruitment At development skills, Rising Up provides reliable behavioral data to guide your HR and educational choices.
.svg){kind=small}
.svg)
.svg)
