The Hiring Process, and Where the Days Actually Go

Search the hiring process and you get the same article fifteen times. Seven steps, eight steps, nine steps. Define the role, write the job description, source, screen, interview, check references, offer, onboard. Every version is correct and not one of them tells you the only thing an operator needs, which is how long each step takes and which step is eating the week.

That omission is not an accident. Stage durations are hard to publish because most teams cannot produce them from their own systems, so the genre settles for the sequence and leaves out the clock.

This piece puts the clock back in. Below is the standard process with a published duration wherever one exists, an honest blank wherever one does not, and the source of delay for each. Then a short method for timing your own, which needs a fortnight and no new tooling.

What is the hiring process?

Two framings are worth separating before the durations make sense. The process is the sequence of stages, roughly the same at a fifty person company and a fifty thousand person one. The timeline is what happens to a specific req as it moves through that sequence. It is the thing hiring managers complain about and the thing good candidates drop out of, and it is where almost none of the published guidance goes.

One number sets the outer bound. Greenhouse's March 2026 benchmark report, drawn from more than 6,000 organisations and 640 million applications, puts median time to fill at 56.7 days, up 37% since 2022. That is the whole process, req open to filled. The question is where those days sit.

The eight stages of the hiring process, and what each one costs in days

Each stage below carries three things. What it is, how long it takes according to published research, and what actually creates the delay.

1. Define the role and approve the req. Duration, no published benchmark. Source of delay, the approval chain and an unwritten scorecard. A req that opens without agreement on what good looks like pays for that later, in a debrief where two interviewers are grading different jobs.

2. Source and collect applications. Duration, no published benchmark, because it depends entirely on channel and market. For scale, Employ's 2026 hiring benchmarks report, covering 6,640 customers, puts average applications per role at roughly 262, rising to 369 in software and tech.

3. Review applications and run the recruiter screen. Duration, 6.6 days average time to screen across that same Employ dataset. Enterprise is fastest at 5.7 days and software and tech slowest at 9 days. Source of delay, the review queue rather than the conversation. The screen itself is thirty minutes.

4. Schedule the first interview. Duration, a median of 6 days from availability request to interview start, per Ashby's recruiting coordination benchmarks across 2.8 million candidates and more than 5.1 million interview events. The same dataset records that 50% of candidates provide availability within 24 hours. Source of delay, therefore, is not the candidate. It is the interviewer calendar on the other side of the request.

5. Run the interview loop. Duration, 22.9 days for the average overall interview process in the US, from Glassdoor's study of 344,250 interview reviews. That study was published in 2015, which makes it the oldest figure here and worth treating as directional rather than current. Source of delay, panel assembly. In the Ashby data, 34% of panel interviews are booked within a day against 54% of single event interviews, so every additional interviewer on a loop is a scheduling cost as well as a calendar one.

6. Collect feedback and debrief. Duration, a median of 2 hours from interview end to scorecard submitted, with 84% to 89% of scorecards in before a candidate advances, again from Ashby. Source of delay, and this is the interesting part, is not the form. Scorecards land fast. What has no published number is the gap between the last scorecard arriving and a decision being made, which in most teams is where the days actually go.

7. Extend the offer and get to a decision. Duration, no published benchmark for elapsed time. Employ puts overall offer acceptance at 81.5% to 89.8%, but acceptance is an outcome, not a clock. Source of delay, compensation approval and the number of people who have to sign.

8. Background check and start date. Duration, no published benchmark, and largely outside your control. Source of delay, notice periods and vendor turnaround.

Three stages out of eight have a number

Look back at the list and the pattern is hard to miss. The three stages with numbers are the three that sit in the middle of an applicant tracking system, where a timestamp gets written whether anyone intended it or not.

The five without numbers sit at the ends and in the handoffs. Req approval happens in an email thread. The debrief decision happens in a meeting. Offer approval happens in a spreadsheet. None of those write a timestamp into the system that reports on hiring, so no benchmark can be built from them, and no vendor can tell you what normal looks like. That is why the steps articles stop at the sequence.

The arithmetic that does not close

Add the three measured stages. Screening at 6.6 days, first interview scheduling at 6 days, the loop at 22.9 days. That is roughly 35 days against a 56.7 day median time to fill, so the difference is about three weeks.

Two caveats before anyone puts that in a board deck. The figures come from three different datasets, gathered in different years, from different populations of company. And they run on different clocks, because time to fill starts when a req opens while the stage figures are candidate level. You cannot subtract one from the other and call the remainder a measurement.

What you can read is the direction. The instrumented middle accounts for roughly two thirds of the elapsed time. The uninstrumented ends and handoffs account for the rest, and they get the least attention precisely because no report surfaces them. A team that spends a quarter shaving days off the interview loop is optimising the part that already has a number.

How to time your own hiring process in a fortnight

You do not need a project for this. You need one export and an afternoon.

Pull your last twenty filled reqs. Twenty is enough to see a shape and small enough that you can fix the gaps by hand.

Record seven timestamps for each. Req approved. First application received. Recruiter screen completed. Availability requested. First interview held. Offer sent. Offer answered.

Use medians, not averages. One executive search with a six month timeline will drag an average into fiction. The median tells you what a normal req does.

Treat the missing ones as the result. If you cannot produce the date a req was approved, or the date an offer was verbally agreed, that is not a failure of the exercise. It is the answer to why nobody has ever been able to tell you where the time goes.

Compare the gaps, not the stages. The durations inside stages are mostly meetings, and meetings are short. The gaps between stages are queues, and queues are where the weeks accumulate.

Run that and you will have something the published benchmarks cannot give you, which is your own eight stage timeline with your own numbers on it. Our hiring process templates cover the stage definitions if you need a starting structure, and the time to hire breakdown goes deeper on the headline metric.

What actually moves the number

Once you can see your own gaps, the fixes are unglamorous and mostly structural.

Name an alternate for every interview stage. Not a person per interview, a second name per stage. When the only qualified interviewer declines, a loop with no alternate moves the candidate rather than the interviewer.

State a weekly interview limit per interviewer. Load is only manageable against a ceiling. Without one, panels fill to the point of strain and the strain resolves itself late in the week, as a dropped slot.

Put a deadline on feedback, before the next round can be booked. The cheapest change on the list, and it attacks the one gap nobody has a benchmark for.

Give the space between stages an owner. Most processes have an owner for every stage and nobody accountable for the handoffs, which is exactly where the unaccounted weeks live.

None of that asks anyone to work faster. It asks the process to stop waiting on one person's calendar.

Frequently asked questions

How long should the hiring process take?

The current median is 56.7 days from req open to filled, per Greenhouse's March 2026 benchmark report. That is a median across more than 6,000 organisations and every role type, so it is a reference point rather than a target. A better test is whether your own median has moved over the last four quarters, and in which direction.

What is the longest stage in the hiring process?

Of the stages with published data, the interview loop is longest, at an average of 22.9 days in Glassdoor's study of US interview reviews. But the honest answer is that the longest stage in most processes is one that nobody times, usually the gap between the last interview and a decision.

Why does scheduling one interview take about a week?

Because the constraint is not the candidate. Ashby's coordination data records a median of 6 days from availability request to interview start, while half of candidates provide availability within 24 hours. The remaining days are spent finding a slot that works for every interviewer on the panel.

How many interviews are in a typical hiring process?

Greenhouse's March 2026 benchmark report puts it at 22.7 interviews per job and 12.3 interview hours per hire across North America. Interviews per hire rose 4.5% and interview hours 5.7% over the three years to 2025, so loops are getting slightly longer rather than shorter.

What is the difference between time to hire and time to fill?

Time to fill measures from req approval to offer acceptance, so it includes sourcing. Time to hire measures from a candidate entering the pipeline to their acceptance, so it measures the process a specific person experienced. Mixing the two is the most common reason two teams report very different numbers for the same hiring.

Can I benchmark my hiring process without new software?

Yes, for the three instrumented stages. Any applicant tracking system will give you screen, schedule and interview timestamps for your last twenty hires. The stages at the ends of the process, req approval and offer decision, usually need a manual pass the first time, which is itself a useful finding.

The bottom line

The steps in a hiring process are not the interesting part. Everybody runs roughly the same eight, and knowing the sequence has never made anyone faster.

The interesting part is that only three of those eight have a published duration, those three add up to about two thirds of a median time to fill, and the remaining three weeks sit in handoffs that no report in your stack currently surfaces. You can find yours with one export and an afternoon.

Start with the number you can already see. Then go looking for the ones you cannot.