Why engineering hires are expensive to get wrong
Every hire is a bet. With engineers, the stakes are higher than most roles, for three reasons. The search is long, because good senior engineers are hard to find. It takes months before anyone knows for sure whether the hire is working, because real productivity comes after ramp-up. And an engineer's work stays behind after they leave: code that someone else now has to understand, fix or replace.
Most companies only count the first and most visible cost, the recruiting. That alone isn't small: SHRM's benchmarking put the average cost per hire across all roles at nearly $4,700 in 2022, and senior technical roles typically cost far more to fill. But it's the costs nobody invoices that make a bad hire painful.
Where the money actually goes
1. Finding them
Recruiter or agency fees, job board spend, and the hours your engineers and managers spend reading résumés and interviewing instead of building. For a senior role, the interview loop alone can cost a team weeks of focus.
2. Waiting for them
Every week the seat is empty is a week the roadmap slips or someone else covers the work. Then notice periods add more weeks before the person even starts.
3. Ramping them up
Even a great engineer needs time to learn your codebase, your systems and your business. During ramp-up they're paid in full and produce a fraction of what they will, and a senior engineer on your team spends part of their week answering questions.
4. Carrying them
When the hire isn't working, the cost spreads. Reviewers spend longer on their pull requests. Deadlines slip. Teammates quietly redo work, and good people get frustrated. This is the most expensive part, and the one nobody tracks, because it shows up as everyone else being slower.
5. Undoing their work
Code written by someone who didn't understand the system has to be found, understood and fixed. In the worst cases, a security shortcut or a badly designed data model stays in production long after the person has gone.
6. Ending it
Performance plans take months. Then there may be severance, the time HR and leadership spend on the exit, and the effect on the rest of the team. Because ending a permanent hire is so painful, many companies wait far too long to do it.
7. Starting again
And at the end of all that, you're back at step one: a new search, a new wait and a new ramp-up, with the roadmap further behind than when you started.
Run the numbers for your own team
There's no single figure for what a bad hire costs, because it depends on the role, how long it takes to notice and how much the team carries in the meantime. You can estimate your own, though. Ask:
- What did the last senior search cost in fees, ads and interview hours, and how many weeks did it take?
- How many weeks does a new engineer on your team take to ship work on their own?
- How long did it take, the last time a hire didn't work out, from the first doubt to the exit?
- Whose time went into reviewing, fixing and covering in that period, and what didn't get built because of it?
- How long did the replacement take to find and ramp up?
Add those up for one bad hire and compare it with the cost of the role for a year. For most teams, the answer is the best argument for taking vetting and the first month seriously.
How to lower the risk of a permanent hire
- Define the role before you hire for it. Write down what the person will own in their first six months. If you can't, it may not be a permanent role yet.
- Assess the actual work. A realistic project built from the role predicts performance far better than puzzles. Our guide to vetting a remote engineer covers how.
- Check identity and history. Fake and misrepresented candidates are a growing, documented problem for remote roles.
- Plan the first 30 days. A structured start surfaces problems in weeks instead of months. See onboarding an embedded engineer.
- Decide early. Set clear expectations for the first months and act on what you see.
When a permanent hire isn't the right bet
Sometimes the lowest-risk move is not to make a permanent hire at all, at least not yet. If the work may not last, if you need a skill you've never hired for, such as your first AI engineer, or if you need someone productive in weeks rather than months, an embedded engineer moves most of the risk off your side of the table.
| Permanent hire | Embedded engineer from us | |
|---|---|---|
| Recruiting fee | Yours to pay | None |
| Time to start | The length of the search, plus notice | Usually within two weeks of agreeing the role |
| Vetting | Your process | Fit, fraud, skills, senior interview and background checks, then your interview |
| If it isn't the right fit | Performance plan, exit, new search | We replace them |
| Ending it | Severance and an exit process | Month to month, with notice agreed up front |
| Employer | You | Sixth Wave |
How our embedded teams lower the risk
Our embedded teams are built around the costs above. There's no recruiting fee and no hourly billing: one monthly rate per person, with onboarding, senior reviews and cover included. Every engineer is vetted for your role and your culture before you meet them, and you interview them before they start. They stay employed by us, so if someone isn't the right fit, we replace them, and engagements run month to month with notice agreed up front.
If the role turns out to be permanent, you've learned exactly what it needs before you hire for it. If it doesn't, you haven't paid for a search, a ramp-up and an exit to find out.