Marcus, this role isn't hard to fill - it's slow to close
Four months open, two declined offers, and a seven-week process. Here's which part of your hiring is losing the people you want, and the order to fix it in.
“Priya was the best backend engineer we've talked to in a year. Six conversations over seven weeks, everyone loved her, and she took something else. She'd had their offer since week two and waited for us out of politeness.”
- Marcus, on the strongest candidate he lost
Where this search is healthy, and where it leaks (higher is healthier)
You scored your panel 4/10 on writing down the same must-haves. A staff spec two people read differently produces six conversations that never converge, which is exactly how seven weeks happens.
Seven weeks and six conversations to a decision. Priya had a competing offer in hand from week two. This is the leak that cost you her, and it is the one you control outright.
Two declines in a row usually means band or level, not persuasion. We can't call it from here - it depends what she was actually holding - but it stops being a coincidence at two.
Nobody left angry, and one of them waited for you on purpose. That's goodwill you earned. What you don't have is anything filling the silences between stages.
The problem is not your pipeline. It's your calendar.
You got the best backend engineer you'd seen in a year to the final stage. That is a sourcing result most teams four months into a search would take gladly, so the story that this role is unfillable does not survive your own answers. What happened instead is that she was in someone else's process at the same time, that process finished five weeks before yours did, and waiting for you was a favor rather than a decision. Underneath the speed sits the reason for it: a panel that scores itself 4/10 on the must-haves cannot close, because nobody in the room knows what would count as enough. Every extra conversation was the panel trying to reach agreement it had never written down. Fix the spec and the calendar shortens by itself.
What to change, in order
This week
Get the must-haves onto one page
- Ask each panel member to write the three things this person must be able to do on day 90, separately, in ten minutes. Compare them in a room. The disagreement you find is the reason this role is four months old.
- Cut the spec to one job. If it names both deep distributed-systems work and mentoring a team, you are hiring two people and interviewing for neither.
- Write down what a 'yes' looks like for each stage, and who owns that call. A stage with no owner becomes another conversation.
Before your next first-round
Compress the process to something a senior candidate can say yes inside
- Collapse six conversations into four stages, and book them as one block with dates offered up front rather than scheduled one at a time.
- Set an internal clock: first conversation to decision in two weeks, and treat a breach of it as a process failure to review, not an unlucky week.
- Name one person who owns the candidate's experience between stages. Silence is where a polite candidate quietly starts preferring the other offer.
- Ask every candidate at first contact where else they are in a process and how far along. Priya's timeline existed from week two - you just weren't holding it.
Before the next search
Settle the offer question so you stop testing it on finalists
- Agree the band and the level for this role internally, in writing, before another candidate reaches the final stage. Two declines is your signal to have that argument now rather than after the third.
- Debrief both declines properly: what they took, what changed their mind, and when the competing offer landed. Ask directly - senior people will usually tell you.
- Decide what you'll flex if the market says no to the band: level, scope, remote terms, equity. Deciding under offer pressure is how good roles get worse.
An honest read on where you stand
Working in your favor
- ✓You are reaching genuinely strong candidates. Whatever your sourcing is, it works.
- ✓Candidates finish your process and stay warm. Priya waited for you, which is not nothing.
- ✓You are asking the question four months in rather than twelve.
- ✓The most expensive leak here - elapsed time - is entirely inside your control, with no budget approval required.
Working against you
- - A panel that scores its own alignment 4/10 cannot make a fast decision, however good the candidate is.
- - Seven weeks means you are, in practice, only able to hire people who have no other option.
- - Two declined offers in a row is a pattern, and it will repeat until someone settles the band and the level.
- - Four months open is also a cost to the team carrying the work, and that cost does not show up in any hiring report.
What hiring managers ask us at exactly this point
Should we just lower the bar?+
No, and your own answers say you don't need to - you found someone you'd have hired. Lower the number of conversations instead. A shorter process with the same bar is a different offer to a senior candidate than a long one, because they are measuring your speed as a signal of how you'll run their first six months.
Is a search fee worth it when we're already finding candidates?+
Not yet, on this evidence. Our contingency search is 20% of first-year base invoiced on the start date, and you'd be paying it to solve a sourcing problem you don't have. What you have is a decision problem, and that is what the Role Clinic and Interview Design are for. We'd rather sell you the $2,400 piece that works than the search that doesn't.
Do we tell candidates the salary range up front?+
Yes, once you've settled it internally. Two declines suggest the band is being discovered by finalists rather than agreed by you, and the finalist stage is the most expensive possible place to find out. Naming it early costs you the wrong candidates in week one instead of week seven.
Bring us the spec and the stage map. Forty-five minutes, no fee.
The Role Clinic is where we read your staff backend spec, your six stages, and your two declines, and tell you which one is actually costing you hires. If the panel still can't agree on the must-haves afterwards, Interview Design is $2,400 fixed and gives you the scorecard, stage map, and interviewer briefs in two weeks. We won't open a search for you until this part is settled.
Four months in, the instinct is to widen the net. Your answers point the other way: you already met the person, and you lost her to a calendar. Shorten the process, write the must-haves down, and settle the band before the next finalist - and the next Priya gets your answer while she's still deciding.