For Indian staffing and recruitment agencies
The same req went to three agencies.Only one shortlist gets paid
It landed at 10:40 on Monday and the client mailed it to two other desks in the same breath. Every CV is scored against the client's real JD, so the five you submit on Tuesday are your five best, not the five you read first.
What an unfilled req actually costs you
The same week, twice
The same req, worked two different ways
You open your Naukri login, run a keyword search, download forty CVs, rename the files and start reading in whatever order they landed in the folder.
Tick applicants on the Naukri, Indeed or Shine tab and they come across at 0.5 credit each. Add the folder of CVs from the weekend drive, a CSV from the old tracker, a LinkedIn profile. One pipeline, duplicates merged into one card.
You submit the five you read first, or the five who picked up the phone. If the client asks why these five, the honest answer is that they were the ones you got to.
The list is ranked against the client's JD. Every profile carries a verdict from strong fit to not fit, matched and missing skills by name, and red flags, so you know what the manager will object to before he says it.
Two evenings on the phone asking the same six questions: current CTC, expected, notice period, buyout, offers in hand, ready to relocate. A third of the numbers do not connect.
An AI voice or video interview in the candidate's own language, one of 11 and more, recorded, with a scorecard and a hiring recommendation sitting on the card. Your calls go to the people worth calling.
CVs reformatted into the agency template and mailed at 11pm. Then three follow-ups, a WhatsApp on Friday, and no idea whether anyone at the client has even opened it.
A teammate checks it first if you want that. It goes out with only the fields you chose to show, and the client's approve, reject or note comes straight back onto the candidate.
An Excel sheet of good people who lost on notice period or CTC, opened once and never again, so next month you pay to source that same market from scratch.
Putting a candidate on another of your own jobs is free, and they are re-scored against the new JD. Email is 1 credit per candidate ever, so every follow-up after the first one costs nothing.
The part only an agency desk deals with
Fill rate and margin, not just placements
An in-house team gets measured on the hire. You get measured on everything you worked and did not bill. Three agencies on one req means two of you paid consultants for a full week and raised no invoice, and that unbilled week is sitting inside the margin on the one placement that did land. So the number that decides whether your desk makes money is not how many people you placed. It is what share of the reqs you accepted turned into a fee, and how much consultant time each fee cost you. Recruitkar is priced around that: credits are spent on work that actually happened, sourcing is refunded when the candidate cannot be found or the match is below the quality threshold, and a candidate you sourced for a req you lost costs nothing to put on the next one.
The reqs you lose sit inside the cost of the ones you win
Sourcing is 1 credit per candidate, refunded when the person cannot be found and refunded again when the match comes back below the quality threshold. A CV screen or a board import is 0.5. A contact lookup is 1 and only charges when a contact is actually found. The searches that came back with nothing come back to you, so the req that died in week two does not quietly bill you for the market you had to read to find that out.
A candidate sourced for a lost req is not a sunk cost
Put the same person on another of your own jobs for free. They are re-scored against that client's JD, so someone borderline on a BFSI req can come back a strong fit on a manufacturing one, and you find that out without paying to source them twice. Email is 1 credit per candidate ever, not per message, so the list you built last quarter stays reachable for nothing.
Ratio is what keeps you on the panel
The first three profiles you send decide whether the client keeps sending you reqs. A ranked, de-duplicated list with missing skills named means you submit your best three rather than your quickest three. And internal review can hold a junior consultant's profile until a teammate approves it, so your biggest account never receives the one that should not have gone.
In practice
How it works on the Monday the req lands
Set the client up once, and keep them apart
Add the client, then the department, then the job. Two clients in the same sector never see each other's pipelines, jobs or boards. Then paste the client's real JD rather than the three lines forwarded on WhatsApp, because everything after this is scored against it, and a three-line JD will come back with empty dimensions instead of a confident number.
Bring candidates in from wherever they already sit
Import the applicants on your Naukri, Indeed or Shine req with the browser extension. Upload the folder of CVs from the walk-in. Drop the CSV out of your old tracker. Pull a LinkedIn profile, share an apply link, or let AI discovery search external profile databases. Six routes, one ranked pipeline, the same person from three sources merged into one card.
Work the top of the list, submit, and see the answer land
Read the verdicts and red flags, then email, WhatsApp or voice-call from the same screen. Send an AI interview before you submit if the client wants proof. Move people across the seven stages from new to shortlisted to hired, have a teammate check a profile first, and share only the fields you choose.
What it costs, in credits
Straight answers
Will this work on my desk?
“I get paid on placements, not on software. Whatever this costs has to come out of a fee I have not raised yet.”
Billing is in credits, and credits go on work that happened: 0.5 to screen a CV or import a board applicant, 1 to source a candidate, 1 for a contact lookup and only when a contact is actually found. Sourcing is refunded when the person cannot be found or when the match is below the quality threshold. The honest limit: a screen or an import is spent whether or not that candidate converts, because the work was done. Only sourcing carries the refund.
“I run two clients in the same sector. If a profile I sent one turns up in the other's pipeline, I lose an account.”
The structure is client, then department, then job, and each client keeps its own jobs, candidates and board. A candidate does not surface inside another client's pipeline unless you put them there, and when you do, that is your decision, with a fresh score against that client's JD. Client sharing also sends only the fields you tick, so a profile can go out without the details you are not ready to hand over yet.
“I have seen matching tools before. Everyone comes back at 80-something and I still read every CV myself.”
That is what the quality gate is for. A two-line JD, or a CSV row that is just a name and an email address, returns empty dimensions instead of a flattering number. You will see fewer scores, not more. The honest limit, and it bites hardest on an agency desk: half your reqs arrive as three lines on WhatsApp. Paste that and you get vague verdicts, and it will tell you they are vague rather than dress them up. Getting the real JD out of the client is still your job.
“Four consultants share this account. One will burn the month's credits by Wednesday and another will mail a half-baked profile to my biggest client.”
Set a per-member credit cap so nobody spends the desk's month on one req. Switch on internal review and a member's candidate cannot reach a client until a teammate approves it, in whichever direction your team works. Nothing leaves for a client on its own. What it will not tell you is which consultant is worth the seat. That is still your read on the floor.
“My follow-up is what closes candidates. If this fires off generic mass mail under my name, it will do more damage than good.”
Then be clear about what the email tool is: a bulk send with merge fields, on a deliverability pool, with bounces monitored so you can see what failed. It is not per-candidate AI-written personalisation, and there is no domain warming service bundled with it. Use it for the wide first touch, at 1 credit per candidate ever so every follow-up after that is free, and keep WhatsApp and a voice call for the people you are actually closing.
Questions
Asked before you ask them
I sourced this candidate for a BFSI client in March and they passed on his notice period. A different client just opened the same role. Do I pay to source him again?
No. Putting a candidate on another of your own jobs is free, and he is re-scored against that client's JD, so someone borderline on one req can come back a strong fit on the next. The database you already paid to build keeps working across your clients.
Two of my consultants worked the same req. One pulled him off Naukri, the other off LinkedIn. Will the client get the same man twice?
No. The pipeline de-duplicates across every source, so those two arrivals merge into one card with one score and one history. One candidate, one submission, and one email credit against him rather than two.
My client sat on five profiles for eight days and then asked why nobody is available. Does this fix that?
Not on its own. Nothing makes a hiring manager reply faster, and it would be dishonest to claim otherwise. What changes is that you can see it happening. The shortlist goes out with approve, reject and note, and the client's answer lands on the candidate card, so you know which of the five have been looked at and which are still sitting there. You chase on day three with a specific name instead of on day eight with a general question.
We ran a walk-in on Saturday and I have 300 CVs in a folder. What happens to them on Monday?
Upload the folder. Screening is 0.5 credit per CV, so the lot is 150 credits, and they come back de-duplicated against everyone already on that job and ranked against the JD, with verdicts, matched and missing skills and red flags. You start Monday at the top of a list instead of the top of a folder.