A bad engineering hire does not just cost their salary. It costs the recruiting fees, the onboarding time, the management attention, the work that needs to be redone, and the months of lost productivity while you find their replacement. The total: $50K-$100K per bad hire. For a startup, that can be the difference between runway and shutdown.
Anatomy of a Bad Hire
How Bad Hires Happen
Bad engineering hires are not always obvious failures. The most common patterns:
The Interview Champion: Excellent in interviews — whiteboard problems, system design discussions, behavioral questions. But struggles with real-world ambiguity, production systems, and collaborative development. Interview performance does not predict job performance.
The Resume Inflator: Strong resume with impressive company names. In reality, they were on a large team where their individual contribution was minimal. Cannot perform independently.
The Culture Mismatch: Technically competent but cannot work in your environment — poor communication, resistant to code review feedback, does not participate in async processes, or creates interpersonal friction.
The Skill Gap: Claimed expertise in your stack but actually worked on different parts of it. "Node.js experience" meant running a script, not building production APIs.
The Timeline of a Bad Hire
Month 1-2: Benefit of the Doubt
- New hire is ramping up. Seems slow but "it takes time to learn the codebase."
- Red flags appear but are attributed to newness.
- Team compensates by doing extra work to cover gaps.
Month 3-4: Recognition
- Pattern is clear. Output is consistently below expectations.
- Code quality requires heavy revision in review.
- Feedback given. Improvement plan created.
- Team morale starts declining (carrying dead weight).
Month 5-6: Decision
- Improvement insufficient. Decision to terminate.
- HR/legal process. Documentation. Severance negotiation.
- Transition plan. Knowledge transfer (of what little they contributed).
Month 7-8: Recovery
- Position re-opened. Recruiting restarts.
- Team fills gap while new search runs.
- New hire found, onboarded, ramped.
- Back to square one — 8 months later.
The Full Cost Breakdown
Direct Costs
| Cost Item | Typical Range |
|---|---|
| Recruiting fees (agency or internal hours) | $15,000-$40,000 |
| Salary paid during employment (5 months avg) | $55,000-$85,000 |
| Benefits during employment | $8,000-$15,000 |
| Severance package | $5,000-$20,000 |
| Equipment (non-recoverable costs) | $1,000-$3,000 |
| Total direct cost | $84,000-$163,000 |
Indirect Costs
| Cost Item | Typical Range |
|---|---|
| Manager's time (interviews, onboarding, coaching, PIP) | 100-200 hours × $100/hr = $10,000-$20,000 |
| Team's time (training, covering gaps, extra code review) | 200-400 hours × $75/hr = $15,000-$30,000 |
| Rework of poor-quality code | $5,000-$20,000 |
| Velocity loss during vacancy + re-hiring | 2-3 months × team impact = $20,000-$50,000 |
| Total indirect cost | $50,000-$120,000 |
Total Cost of a Bad Hire
Conservative estimate: $50,000-$100,000 Realistic estimate (including full indirect costs): $100,000-$280,000
For a startup with $500K-$1M of runway, one bad hire consumes 10-50% of remaining resources.
The Retained Partner Alternative
A retained engineering partner (like Kwiqwork's dedicated team model) eliminates most of these risks:
No Recruiting Cost
- We provide vetted, experienced developers
- No job postings, no agency fees, no 20-interview funnels
- Time to first productive developer: 2 weeks (not 3-6 months)
Free Replacement Guarantee
- If a developer does not perform well, we replace them within 2 weeks
- No severance, no HR process, no awkward conversations
- We absorb the replacement cost — your project continues without interruption
2-Week Paid Trial
- Evaluate fit on a real project before any commitment
- If the trial is not satisfactory, you owe nothing beyond the trial period
- No risk of the 5-month discovery period for a bad hire
No Single Point of Failure
- Our team includes backup knowledge across developers
- If one person is unavailable (sick, vacation), the team compensates
- No scenario where one person leaving means starting over
Risk-Adjusted Cost Comparison
To compare fairly, you need to account for the probability of a bad hire:
Industry data:
- ~20% of engineering hires do not work out within 12 months (various sources estimate 15-25%)
- Average time to identify a bad hire: 4.5 months
- Average time to replace: 3 months
Risk-adjusted cost calculation (in-house hire):
Expected cost = (80% probability of good hire × annual salary) + (20% probability of bad hire × bad hire cost)
For a $170K/year developer:
- Expected annual cost = (0.8 × $170K) + (0.2 × $100K bad hire cost) = $136K + $20K = $156K risk-adjusted
That is 91% of salary as the risk premium just from bad hire probability.
Retained partner cost:
- Monthly cost: $2,000/month = $24K/year per developer
- Bad hire risk: $0 (free replacement)
- Risk-adjusted annual cost: $24K
The risk-adjusted comparison: $156K vs. $24K (6.5× more expensive in-house after accounting for hiring risk).
When In-House Hiring Still Makes Sense
Despite the cost analysis, some roles should be in-house:
Technical leadership: CTOs, VPs of Engineering, Staff Engineers who set architectural direction need institutional context and long-term commitment.
Core IP development: If the algorithm IS your competitive advantage, it should live in heads that are contractually bound to your company.
Regulatory requirements: Some compliance frameworks (SOC 2, HIPAA) prefer or require certain roles to be in-house employees.
Culture building: The first 3-5 engineering hires define your engineering culture. These should be deliberate in-house decisions.
For everything else — feature development, capacity scaling, specialized skills — a retained partner delivers equal or better output at a fraction of the risk-adjusted cost.
How to Minimize Bad Hire Risk (If You Must Hire)
If your situation requires in-house hiring, reduce bad hire probability:
Paid trial project before offer. Pay the candidate for 1 week of real work. Evaluate actual output, not interview performance.
Reference checks with technical peers. Talk to engineers who worked WITH them (not just managers). Ask specific questions: "What did they build? How did they handle disagreements? Would you hire them again?"
Team-based interview. The people who will work with the hire daily should interview them. Cultural fit matters as much as technical skill.
Clear 90-day expectations. Define specifically what "success" looks like at day 30, 60, and 90. Review against these milestones. If day 60 expectations are not met, do not wait until month 5 to act.
Probation period with defined exit. 3-month probation with clear performance criteria and simplified termination if not met.
Even with these practices, some bad hires will slip through. The question is whether the risk premium is worth it for the specific role — or whether a retained partner eliminates the risk entirely for roles where institutional commitment is not critical.