When a client tells us hiring is slow, they almost always blame sourcing. In four years of funnel diagnostics, sourcing has been the binding constraint in fewer than a third of cases. The delay is usually sitting in the two weeks between a shortlist arriving and someone reading it.
The nine interventions below are ordered by what we have actually measured: median days removed from requisition-to-offer, against the internal effort required to implement. Start at the top. The first three are almost free.
The scoreboard
| # | Intervention | Median days saved | Effort |
|---|---|---|---|
| 1 | Pre-book the interview slots | 9 | Low |
| 2 | Collapse the loop to four stages | 7 | Low |
| 3 | Write the scorecard before the advert | 6 | Low |
| 4 | Agree the compensation band up front | 6 | Low |
| 5 | Give a 48-hour feedback SLA | 5 | Medium |
| 6 | Batch the panel, decide same day | 4 | Medium |
| 7 | Replace the take-home with a live exercise | 4 | Medium |
| 8 | Delegate offer approval below CFO level | 3 | High |
| 9 | Keep a warm bench for repeat roles | 11 | High |
1. Pre-book the interview slots
The single largest source of delay in almost every funnel we have examined is calendar friction. A hiring manager who holds two recurring one-hour slots each week loses nothing when there is no candidate, and saves nine days when there is.
Candidates read scheduling speed as a signal of how much you want them. A firm that can interview on Thursday when the shortlist landed Tuesday wins people from firms offering more money three weeks later.
2. Collapse the loop to four stages
Median interview loops in Indian technology hiring now run to five and a half stages. We have never seen evidence that stages five through seven improve the quality of the decision, and we have plenty that they cost candidates.
Four stages is enough for almost every role: a screen, a deep technical or functional assessment, a manager conversation covering scope and trajectory, and a cross-functional or leadership round. Anything beyond that is usually an organisation avoiding a decision rather than improving one.
Every additional interview stage beyond the fourth cost our clients a median of 3.1 days and a 6% drop in offer acceptance.
3. Write the scorecard before the advert
A job description describes a role. A scorecard describes how you will know someone can do it. Four to six competencies, each with the evidence that would prove it, agreed by everyone in the loop before a single CV is reviewed.
The saving is not in the writing, it is in the arguments you never have later, the ones where two interviewers realise in week five that they were assessing different jobs.
4. Agree the compensation band up front
Roles budgeted below the market median took 2.4 times longer to fill in our data. That is not a sourcing problem and no recruiter can outwork it.
Establish the band against live offer data before the requisition opens, and get finance to sign it then. Renegotiating a band at offer stage costs eleven days on average, and by then you have usually lost the candidate who prompted it.
5. Give feedback within 48 hours
An SLA on interview feedback is the cheapest quality improvement available to a hiring team. It removes the dead air where candidates accept other offers, and it forces interviewers to record their reasoning while they still remember it.
Make it structural: feedback is submitted before the interviewer leaves the room, or the slot is not considered complete.
6. Batch the panel and decide the same day
Where the loop involves three or more separate people, run them consecutively on one day and hold a fifteen-minute debrief immediately afterwards. Decisions made in that debrief are both faster and, in our experience, better calibrated than decisions assembled from written feedback over a week.
7. Replace take-homes with live exercises
Take-home assignments have the worst effort-to-signal ratio in the modern loop. Median turnaround in our data is nine days, completion rates sit near 60%, and the strongest candidates, who have options, decline them most often.
A 45-minute live exercise with a real problem produces better signal, no delay, and a conversation the candidate usually enjoys.
8. Delegate offer approval
If every offer routes through a CFO, you have built a queue. Set banded approval limits so a hiring manager and a talent lead can issue anything inside the agreed range without escalation. Only exceptions should travel upward.
9. Keep a warm bench for repeat roles
For roles you hire more than four times a year, maintain a relationship with fifteen to twenty qualified people who are not looking today. It is the highest-effort intervention here and by far the highest return: median eleven days saved, and a materially better hire because you are choosing rather than searching.
Want your funnel diagnosed?
We run a funnel audit for clients and prospects: conversion at every gate, benchmarked against 340 comparable organisations, with a ranked list of what to fix first. It takes about a week and there is no obligation attached to it.
What not to do
- Do not lower the bar. Every organisation that tried this in our sample paid for it within two quarters through regrettable attrition and re-hiring.
- Do not add recruiters to a broken funnel. If the constraint is decision latency, more sourcing produces a larger queue, not more hires.
- Do not measure only time-to-fill. Pair it with 90-day retention, otherwise you will optimise your way into hiring the available rather than the suitable.
