Reducing time to hire: attack the queues, not the interviews
Almost all of the elapsed time in a hiring process is candidates waiting, not candidates being assessed. Teams under pressure to move faster usually cut assessment, which is the small part, and leave the queues, which are the large part, untouched.
- Measure dwell time per stage, not just total days. The total tells you there is a problem; the stages tell you where.
- The two largest queues are almost always application-to-first-contact and interview-to-decision.
- Compressing the interview itself buys days and costs signal — the worst available trade.
- Parallelise assessment, serialise only the decision.
- Every day of delay is a chance for a competing offer, and the strongest candidates have the most of them.
Where the days actually go
Instrument a typical process and the shape is remarkably consistent: assessment consumes hours, waiting consumes weeks.
| Stage | Typical dwell | What is actually happening |
|---|---|---|
| Application → first review | Days to weeks | The pile waits for someone to have time |
| Review → first contact | Days | Scheduling attempts, phone tag, timezone friction |
| First interview → next round | Days to weeks | Panel availability — the hardest queue to shorten conventionally |
| Final interview → decision | Days to weeks | Waiting for the full slate, deliberation, approvals |
| Decision → offer out | Days | Approvals, banding checks, contract preparation |
| Total assessment time | Two to four hours | The only part most teams try to shorten |
The implication is uncomfortable but simple: if you shortened every interview by a third you would save perhaps an hour, and lose the depth that made the interview worth running. If you removed one week-long queue you would save five days and lose nothing.
The first queue is the one that costs you candidates
Application-to-first-contact is both the longest queue in most processes and the one candidates judge you on. It is also where the strongest applicants leave, because they are the ones with other conversations running.
It exists because reviewing applications requires human attention that is rationed. So the queue length is a direct function of assessor capacity divided by applicant volume — which is why the queue gets longer in exactly the rounds where speed matters most.
- Removing the capacity constraint at the first stage is the single highest-leverage change available to most hiring processes.
- Assessment that runs whenever the candidate is available — overnight, between shifts, across time zones — eliminates the scheduling component entirely.
- Continuous shortlist delivery beats batch delivery: send the hiring manager candidates as they qualify, not when the round closes.
- Set a service level for first contact and report against it. Anything unmeasured drifts.
The decision queue, and the slate problem
The second large queue is waiting for a complete slate before deciding. It is defensible in principle — comparative assessment is better than sequential — and it is where good candidates are lost while the process waits for a fourth person who was never going to be better.
- 01Set a slate deadline, not a slate size. Decide from whoever qualified by the date.
- 02Pre-authorise the decision. If the hiring manager needs three approvals to make an offer, the approvals are the process, not a formality.
- 03Separate "is this person hireable" from "is this person the best of the slate". The first can be answered immediately and unlocks a conditional conversation.
- 04Keep the evidence comparable so a fast comparison is possible at all — this is what an anchored rubric with cited evidence buys you.
Why speed and quality are not opposed here
The assumed trade-off — move faster, assess worse — holds only if you shorten assessment to move faster. Attack the queues instead and the relationship inverts, for a specific reason.
A faster process assesses a better pool. The candidates who leave a slow process are disproportionately the ones with alternatives, which is to say the strong ones. Ten days of delay does not just cost you ten days; it changes who is still available at the end of them.