What 70% faster time-to-fill actually looks like

70% faster time-to-fill is what Hyatt measured on high-volume hospitality roles, alongside 7x ROI. A number that size is never produced by people working faster. It comes from deleting waiting — the days a candidate sits between applying and being spoken to, and the days a recruiter sits between finishing screens and having something to hand a hiring manager.

This is a walkthrough of where those days live and what removes them. No customer anecdotes: the mechanics are the interesting part.

Where do the weeks actually go?

Almost never inside the interviews. A screening interview takes twenty minutes; a screening stage takes three weeks. The difference is entirely gaps.

• Apply to first contact. The application lands in a queue and waits for a recruiter to reach that part of their week.
• The scheduling round-trip. Two or three emails to find a slot, then a reschedule, then a no-show that costs the slot anyway.
• Working-hours overlap. A candidate on shift work or three time zones away can only be interviewed during a narrow band that also has to be free on your side.
• The review backlog. Screens completed on Tuesday sit unsummarised until Friday, because writing up twelve conversations is nobody's favourite afternoon.
• The handoff. The shortlist waits for a debrief meeting that exists because nobody trusts a document to carry the reasoning.

Four of those five are dead time. None of them is anyone's fault, and none of them shows up in a metric called "interviewer capacity."

What does 24/7 screening remove?

The scheduling round-trip and the working-hours constraint, which between them are usually the largest block. The candidate picks a slot and takes the interview then — there is no negotiation because there is no second calendar to satisfy.

In practice that means a candidate who applies at 2am gets the first conversation before your team is awake, and a nurse coming off a night shift interviews at 11pm on her phone, in the language she thinks in. Neither of those interviews was competing for a slot with anything. The interval that used to be measured in weeks becomes an interval the candidate chooses, which is a different kind of quantity.

What does the recruiter open on Monday?

A ranked shortlist with the reasoning already attached. Every candidate carries a score against the same rubric, and every score points to the transcript moment that earned it.

That last property is what kills the second delay. A shortlist a hiring manager can audit — read the number, read the sentence behind it, disagree in writing — doesn't need a debrief meeting to be actionable. The review stage stops being a queue of things to watch and becomes a list to read: a change in the unit of work, not just its speed.

What happens to the review backlog at volume?

It stops being the constraint. Gammon Construction reached shortlist 40% faster while screening thousands of applicants without adding headcount, which is the specific claim worth noticing: the volume went up and the team didn't.

This is where formats diverge sharply. Recorded video removes the scheduling cost and leaves the review cost in place, so at a thousand candidates you've converted a scheduling problem into a watching problem and handed it to the same people. A scored interview removes both, because the artefact reaching the reviewer is already ranked and already explained. The comparison in full: AI interviewer vs. one-way video interview.

Which stage becomes the new bottleneck?

The hiring manager's calendar, almost immediately. Once screening stops rate-limiting the funnel, the next constraint is a human one and it appears within the first fortnight.

Teams that are ready for this handle it by holding dual availability at shortlist time, so the manager's next slot is tentatively placed before anyone has read the list. Teams that aren't ready discover that they've moved the queue rather than removed it, and conclude the tool didn't work. It's worth deciding in advance which failure you'd rather have, because a faster screening stage feeding an unchanged decision loop produces a visible pile-up.

What doesn't get faster?

Everything downstream of the human decision, and it's worth being explicit so the business case doesn't overreach.

1. Final-round panels. Real people, real calendars, and rightly so.
2. Offer approval and compensation review. Internal process, unchanged.
3. Background and reference checks. Third-party clock.
4. Notice periods. No software touches this, and in senior hiring it's often the largest single block in the whole cycle.

Nor does speed substitute for a rubric. A funnel that moves faster against a vague standard just reaches the wrong shortlist sooner. The screening stage is only worth accelerating once you can say what it's screening for — how to write interview rubrics an AI can score consistently.

How do you tell whether your funnel has this problem?

Four measurements, none of which require a vendor.

• Median days from application to first conversation. If it's over five, the gap is where your cycle time lives.
• Share of screens that took more than two emails to book. This is the coordination tax, and it's usually larger than anyone guesses.
• Drop-off before the first conversation. Candidates lost here don't appear as losses, only as a smaller funnel.
• Distribution of interview start times. If every interview happened between 10am and 4pm on a weekday, you've been selecting for availability without meaning to.

Common questions

Is 70% a typical result?

It's Hyatt's result on high-volume hospitality roles, and we quote it as an outcome rather than a promise. Gammon's was 40% faster to shortlist. Your number depends almost entirely on how much of your current cycle is waiting rather than working — funnels with little dead time have little to reclaim.

Does moving faster mean screening less carefully?

The rubric doesn't change with the clock. The same criteria are applied to every candidate with 100% consistency, and the speed comes from removing coordination rather than from shortening the conversation.

What has to be in place before any of this shows up?

A rubric the hiring manager agrees with, and the ATS connection so shortlists land where recruiters already work. Integration is typically a one-day job across 30+ systems; the rubric takes longer and matters more.

If you want to know where your own weeks are going, the shortest route is watching an interview run end to end against a live req. We'll set one up.