The HireQwik High-Score Line That Pings Slack

A high score candidate alert threshold is the line between a channel that helps and a channel people mute. I watched both failures on the same fresher role, two weeks apart. In week one the line sat at the default and the pod got dozens of pings before lunch. They muted the channel. In week two someone, stung by the noise, moved the line so high that the three people they later hired never produced a ping at all. The hires were found on Friday, inside the inbox, by a recruiter who did not trust chat anymore.

The threshold did not score anyone. It only chose who was allowed to interrupt. That choice is easy to get wrong if you treat 0.85 as a universal definition of “top candidate.”

What the HireQwik high-score alert actually measures

The ping runs when the sheet poller inserts a resume row. The payload is the match score on that row. HireQwik compares it with the organisation’s threshold. Leave the box alone and the cut is 0.85. Clear the cut, with the event ticked, and chat gets one line: who matched, the percent, which opening, and a URL into that resume record.

Under the cut, chat says nothing. The person is still in the inbox. Quiet does not mean rejected.

The event itself starts off. You can leave 0.85 stored for months and never send this note, because the tick is separate from the number. I would rather you catch it mid-week, well before a college sheet hits at dawn.

The input accepts a fraction. Its label says the range is zero through one, in steps of a hundredth. 0.85 means 85 percent. 1 means every point. 0 lets every ingested score through, which is a firehose. When you leave the field, an empty or junk value is pulled back inside that range. Do not half-type and assume the save became 85.

This is not the voice verdict. Resume match and the interview verdict answer different questions. The alert you are tuning here is the first of those questions: did the paper clear a bar worth interrupting a human?

Why the default cut is not a hiring bar

The inbox already sorts resume matches into four bands. The top band opens at 90 percent. The band under it opens at 75. Maybe opens at 60. This alert is not a fifth label. Leave the default untouched and an 86 still interrupts, because 86 percent clears 0.85, and a 90 does too. A 74 stays quiet, even when a bunched campus rubric makes 74 a serious student.

That is a reasonable default for a role where the score spread is wide and an 85 is rare. It is a poor default for a role where serious applicants cluster in the high 80s, and a poor default for a role where you only want the 90s in chat.

Look at one finished day’s histogram before you arm the event. If the people you would have phoned sit at 78 to 84, a default line hides them. If half the sheet sits above 85, the default line is not a filter. Move the number, or do not turn the event on.

Becker, Connolly, and Slaughter studied 3,012 hires at one company and found earlier offers were more likely to be accepted, with no performance or turnover penalty for moving faster. The paper is from 2010. I am not pretending it is a 2026 India campus study. I am using the narrower point: speed after you already know someone is strong is not free to waste. A threshold that hides the strong ones wastes it. A threshold that cries wolf trains the team to waste it too, because they stop reading.

When the HireQwik line floods a fresher drive

A loose fresher rubric plus 0.85 plus a 1,000-row night is how a channel dies. Each qualifying row is its own message. There is no rollup inside this event. Fifty clears means fifty notes.

If that is your sheet, do not “solve” it by muting. Pick one of these:

• Raise the line to 0.90 so chat only hears the resume Strong Go band, and let Go wait for the morning list.
• Turn the instant event off. Keep the morning email to the role owner. On days with new work it arrives at nine, India time, with a count of fresh profiles and a split of the two positive interview tiers, plus a link to that role’s queue. Chat stays usable.
• Keep the instant event for one specialist role, and keep it disabled on the volume role. The threshold is organisation-wide, so you cannot set 0.90 for freshers and 0.75 for a lateral req on this single box. If those roles need different lines, the instant ping should follow the stricter role, and the looser role should live on the digest.

The organisation-wide part surprises people. Sinks and the high-score number are not per job description. Accept and reject automation is chosen on each job description. Do not assume the alert line follows the same map. When one workspace holds both openings, they share this cut.

When a higher line hides the people you will hire

The opposite mistake feels responsible. After a noisy day, a lead sets 0.95 or 0.98 so the channel “only gets the obvious ones.” On many rubrics the obvious ones are a handful, and the people you hire sit just under the new line with a clean voice screen later.

A resume at 0.88 sits in the second resume tier. At a 0.95 cut, chat stays quiet for that person. It can still earn a top interview verdict once they have talked. If your process depends on chat to notice Go resumes, a vanity line creates a silent miss. The inbox did not miss them. The habit of not opening the inbox did.

I would rather the channel hear a slightly wider set for a week, while someone checks whether the extra names were junk, than hear a perfect trickle and discover the hires in a Friday export. Write down the line and the date you set it. When a manager later asks why an 0.87 never appeared in Slack, you can point at the number instead of at a memory.

The HireQwik digest nudge is the pile alarm

Instant messages are bad at “you now have a queue.” They are good at “this one person just cleared.” The pile has its own trigger.

After the last note, if still-undecided outcomes in the top positive tier and the tier under it add up to five, HireQwik emails a nudge. Maybes are not in that sum. A role can change the five. Mirroring it into chat is opt-in, on the same switch as the daily roundup. I like the nudge in the mailbox even when I hate instant pings, because it speaks as a count. “Six strong candidates are waiting” is actionable. Sixty individual stars before noon is not.

The nudge is about interview verdicts waiting for a human, not about the resume threshold. You can run a strict high-score line and still want the nudge, so a top interview outcome does not sit untouched just because the resume ping was the only habit you built.

On Hold still needs a deadline if your queue’s real risk is the middle, not the top. The nudge will not nag you about On Hold. It watches Strong Go and Go. If your team parks everyone in the middle, this alarm stays quiet and you have a different problem.

How I would set the line for two kinds of role

Volume campus trainee. Do not arm the instant high-score event at 0.85 until you have seen the spread. If the spread is tight and high, either move to 0.90 or skip the instant event. Turn on the morning digest email. Leave the nudge at 5 unless a single day regularly produces more unreviewed Go results than a pod can clear. Put Candidate screened on only if someone will act the same day.

Scarce lateral role, a few dozen applicants. 0.85 is often fine, and sometimes low if you would happily read a 0.80. Because you cannot give this role its own number, decide whether the shared line should serve this role or the volume role. If the volume role would flood at the lateral role’s line, keep the instant ping off and let this recruiter live in the inbox filter. The score filter on Needs Review is the tool for “show me this band,” and it does not interrupt anyone.

In both cases, run Send test after you save, then watch one real row cross the line so you know the percent in the message matches the inbox. The silent webhook failures are a different bug: right people, no delivery. This page is the bug where delivery works and the wrong people are selected.

What the alert text contains and the Teams address to paste are the setup around this number. The number itself should be chosen from a score distribution you have seen, not from the default that shipped because a box cannot be empty.

If you want another read on a real histogram before you arm the ping, use the demo form. Bring a day’s score spread with names removed. The line is a small field with a large effect, and it is cheaper to set it once than to retrain a channel that learned to mute you.