That said, I've also given developers another week to wrap things up, more time because of illness and family obligations, and a couple more weeks to start when they asked because "they were still burnt out".
It's not like I throw my hands up and the first road block and shred the resume.
If the person wants to stay to finish one thing -despite the terrible workplace-, that's actually a major display of [misplaced] commitment.
That's only half the coin - the other half is "would you rather restart the hiring process than hire someone showing red flags (not sticking to their commitments / showing up on time) before they even start the job?"
Sick? Family emergency? Funeral? Sure.
Can't let go of your old job? Hmm...
> From what I've read about the numbers this seems very costly.
From what I've read, so is hiring the wrong person and then having to fire them or have them quit - all that time wasted getting them up to speed.
On the other hand, they may be getting cold feet about leaving all their old colleagues and friends behind and about to accept a generous counteroffer that they haven't yet had time to formalize.
Or perhaps they remembered they really hate PHP. Or the state they'd be moving to. Or gotten a last minute job offer from someone else that they like way better. Or...
Plus, I'm always concerned when a developer ends their last job on Friday and starts Monday at our company.
First several weeks will be ramp-up anyway.
If I want to avoid coverage loss, I'll go with back-to-back employment rather than risk a few days and having to enroll in COBRA.
I'm more concerned about the scenario where a developer has been working 12+ hour days on a release and then 50 or so hours later is now trying to ramp up on at our company.