A Formula 1 pit crew is judged on one number: the time between the car stopping and the car leaving. Two seconds is good. Under two seconds is exceptional. Teams spend millions shaving tenths off that number, because in racing, speed and time is the entire point.
But every pit crew also knows the one shortcut they can never take. Every wheel nut must be torqued to spec before the car is released, because a wheel that comes loose at 300 kilometres an hour does not just cost the race. It costs the driver.
So, the crew optimizes everything except that one check. They get faster at everything around it, never through it. Niyam, Orbrick’s VP of HCM and Design-Thinking, talks about something similar in his blog What gets Measured, gets Gamed.
Recruiting teams do not have that discipline. Time-to-hire became the metric everyone tracks, everyone reports upward, and everyone gets rewarded for shrinking, without ever asking which check quietly got skipped to make the number smaller.
Here is the uncomfortable question this piece is built around: what did you actually cut to hire this fast, and have you ever gone back to check if it mattered?
Are your fastest hires becoming your fastest exits?
Start with the number your recruiting dashboard already treats as a win. Time-to-hire, measured from requisition opened to offer accepted, sitting in your quarterly report as evidence the function is running well.
Now pull a different report. Take every hire from the last 18 months and split them into four groups based on how fast they moved through the pipeline. Fastest quarter, second, third, up to the slowest quarter. Against each group, lay their first performance rating and whether they were still with the company at the 12-month mark.
In most organizations that run this check for the first time, a pattern shows up that nobody in recruiting was looking for. HRD America’s article published in 2025 shows that the faster recruitment processes lead to lower efficiency and higher turnover rates.
That is not a coincidence, and it is not a reflection of those candidates being worse employees. It is what happens when the process that would have caught a mismatch gets shortened to hit a number that was never actually tied to quality in the first place.
What did you actually cut to get faster?
This is where most recruiting reviews stop, because the correlation alone is uncomfortable. But a correlation without a mechanism is just a statistic someone will argue with in the next leadership meeting. You need to show exactly what got sacrificed.
Pull the stage history for every requisition in that fast quartile. Count how many interview rounds actually happened against the standard defined for that job family. Check whether the skills assessment ran at all or was marked optional and quietly skipped. Check whether the “panel interview” was actually a panel, or one person on a call doing the job of four.
The pattern is rarely random. One stage tends to disappear first, and it is almost always the one most dependent on scheduling multiple people at once. The technical assessment. The structured panel. The reference check that becomes a box someone ticks without actually dialing the number.
That stage was in the process for a reason. Someone designed it to catch something a resume and a single conversation cannot. Removing it does not remove the risk it was built to catch. It just removes your ability to see that risk before the offer goes out.
Who is making the trade-off, and do they even know it?
Here is the part that turns this from a process problem into a people problem, and the part most recruiting leaders would rather not look at directly.
Segment the hires which had the lowest time to hire and exited the organization by the 12-month mark by the requisition owner. Not by department, by the actual recruiter or hiring manager who ran the process.
If the pattern spreads evenly across the whole function, you have a systemic design flaw, and the fix is a process redesign. But that is rarely what the data shows. Usually, it concentrates. A handful of recruiters under headcount pressure, or a couple of hiring managers who treat an open seat as an emergency every single time, account for most of the fast-and-risky hires.
Ask them why, and the answer is almost never “I don’t care about quality.” It is closer to “the team was drowning, I needed someone in the seat, and I figured we’d course-correct in the first few months if it didn’t work out.”
That course-correction rarely happens the way people imagine. By the time performance issues surface, the new hire has already onboarded the rest of the team into their way of working, the manager who rushed the hire is now invested in defending the decision, and letting the person go quietly becomes more expensive, politically and financially, than it would have been to just run the extra interview round.
Summing up
Organizations that get this right stop treating time-to-hire as a standalone number and start treating it as one half of a pair. Speed only means something next to the outcome it produces. A fast hire who performs and stays is genuinely a win. A fast hire who exits in month four was never a “win”, just a cost that arrived later, in the form of rehiring and training.
The fix is not to slow every requisition down across the board. Urgency is sometimes real. The fix is to know, before you cut a stage, whether that stage was actually catching something, and to have the discipline to protect the two or three checks that matter even when the pressure to fill the seat is highest.
Oracle Recruiting Cloud already has everything needed to run every check in this piece. Stage timestamps, interview panel composition, assessment scores, and offer decisions all sit in the requisition record. Performance Management holds the outcome data on the other end. The join between the two takes an afternoon for anyone who decides to run it, yet almost no organization has connected their time-to-hire dashboard to their performance-outcome dashboard, because they live in different reports, owned by different teams, reviewed in different meetings.
This is exactly the kind of gap an AI agent should be watching instead of a person remembering to check quarterly. A custom workflow agent, built within Oracle AI Agent Studio, that flags any requisition where time-to-hire drops below a set threshold and a standard stage was skipped, routing it for a manual quality review before the offer goes out rather than an exit interview eighteen months later, closes that gap without slowing down the hires that genuinely do not need the extra step.
There’s no universal number for how fast is too fast, and that’s part of the problem. Only one in five organizations even measures quality of hire, so most companies are optimizing time-to-hire with nothing to check it against.
Speed was never the problem. Speed without knowing what it cost you is.