Ask a team their time-to-hire for engineers and you get a number that mixes "days since we opened the req" with "days since we started seriously looking." Those are different clocks. This guide splits the timeline so you can see which stage is actually burning weeks.
A realistic breakdown
For a mid-level IC role with a working process:
| Stage | Healthy duration | Where it explodes |
|---|---|---|
| Scorecard + salary lock | 1–2 days | Debating requirements for three weeks |
| Sourcing shortlist | 3–7 days | Searching without filters or proof |
| Outreach → first reply | 3–10 days | Generic emails, no salary |
| Screens | 3–5 days | Calendar chaos across timezones |
| Deep interviews / exercise | 5–10 days | Take-home with no deadline or review SLA |
| Decision + offer | 2–4 days | "Let's sync next week" loops |
| Notice period | 2–12 weeks | Not under your control |
If your "time to hire" is 90 days, notice period may be half of it. Fix the controllable middle first.
The three leaks that add the most days
1. Sourcing without a scorecard
Starting outreach before you agree on must-haves produces the wrong shortlist, which produces restarts. Lock the scorecard first — stack, seniority, timezone, salary — then source. Remote hiring sequence covers the same discipline.
2. Outreach that does not earn replies
Every day of silence is usually a copy problem, not a market problem. Specific, short, salary-transparent notes outperform sequences. Templates: how to email developers.
3. Decision lag after the last interview
The candidate is done; your team is "waiting for alignment." This is the most expensive delay because strong candidates are also talking to someone faster. Rule: same-day debrief, offer decision within 48 hours of the final round.
A 21-day controllable target (excluding notice)
| Days | Milestone |
|---|---|
| 0–1 | Scorecard, range, interview panel locked |
| 2–6 | Source and screen 30–40 profiles on proof (GitHub screen) |
| 5–8 | Outreach wave 1 (15 people) |
| 8–12 | Screens + exercise sent |
| 12–18 | Final interviews |
| 18–21 | Offer out |
Miss the window? Restart sourcing in parallel on day 14 — do not wait for the first wave to fully die.
What not to cut
A real look at their work. Skipping repository or work-sample review to "go faster" creates failed hires, which is the slowest outcome of all.
Reference or background checks your company requires. Compress calendar, not compliance.
Candidate questions. Slow answers here feel like disinterest and kill acceptance rates.
Tooling that actually shortens the clock
- Faster shortlists: filters on activity and stack beat manual LinkedIn paging — start from hire-developers.
- Fewer no-shows: calendar links with timezone detection.
- Fewer loop-backs: written scorecards shared with every interviewer before the call.
- Fewer stalled offers: pre-approved bands so comp does not restart negotiation internally after the candidate expects an offer.
Buying another ATS will not fix a team that cannot decide. Buying sourcing help will not fix a three-week take-home nobody grades.
How to measure it next quarter
Track separately:
- Days from scorecard-ready to first outbound.
- Days from first outbound to first onsite/final.
- Days from final to offer sent.
- Days from offer to start (includes notice).
Improve (1)–(3). Report (4) honestly so leadership stops blaming recruiting for resignation clocks.
If (1) is the problem, you have a sourcing issue — free sourcing or a specialist tool. If (3) is the problem, you have a leadership issue — no software fixes that.