How to Track Candidates and Placements for Recruitment Agencies (2026 Guide)

A recruitment agency is two businesses stacked on top of each other. The first is a sales business: pitching clients, winning job orders, and getting paid on placements. The second is a relationship business: keeping a pipeline of candidates warm enough that when a job order lands on Tuesday, you have three submissions in front of the hiring manager by Friday. The agencies that scale past one or two desks are the ones that treat candidate and placement tracking as the operational spine of both businesses — not as a Trello board the recruiters update when they remember.

This guide walks through how to set up that spine in 2026: the ATS structure, the 6-stage candidate funnel, job orders, submissions, interview coordination, the placement record with commission, and the guarantee window that decides whether the placement actually counts. It is written for agency owners, billing managers, and senior recruiters at agencies between two and fifty desks who are either replacing a spreadsheet-and-email setup or rebuilding one that grew out of control.

The goal is a system where every candidate has one record, every job order has one record, every submission ties a candidate to a job order with a date and a fee, and every placement has a start date, a fee structure, and a guarantee end date that the system surfaces automatically. If your current setup cannot answer the question 'how many candidates are at first-interview stage across all open job orders right now,' the steps below are how you fix that.

Every recruitment agency has a funnel, whether or not it is written down. The agencies that consistently bill through the funnel rather than around it use six stages — six is enough to model the work without splintering reporting into meaningless slivers. Each stage is a status on the candidate-job-order pairing (the submission), not on the candidate alone. A candidate can be at different stages on different job orders simultaneously, and your system must handle that without forcing a single global status.

Before any candidate or job order goes in, the system needs three structural decisions made and locked: the candidate record schema, the company and contact schema, and the job-order schema. Most agency tracking pain comes from changing these structures three months in, when 800 candidates are already in the system with inconsistent fields.

For candidates, the minimum viable schema is: full name, primary email, primary phone, current title, current employer, location, work authorization status, total years of experience, target salary range, notice period, source (where they came from), owner (which recruiter), tags (skills, certifications, industries), and resume attachment. For company contacts: company name, contact name, role, email, phone, decision-making authority (recruiter, hiring manager, HR, executive), and the agency relationship terms (fee percentage, exclusivity, payment terms, guarantee period). For job orders: client company, hiring contact, job title, location, employment type (perm, contract, contract-to-hire), salary range, fee terms, opened date, target close date, status, and assigned recruiter.

A platform like [Deelo's CRM](/apps/crm) lets you define these schemas with custom fields against the standard Contact and Company objects, then layer Practice/Matters on top to model job orders and placements as their own first-class records. The advantage of building it this way rather than buying a single-purpose ATS is that the same system handles your business-development pipeline (prospect clients, pitches, signed agreements) in the same database — recruitment agencies that separate BD from delivery into two systems lose the obvious feedback loop between them.

The candidate record is the asset. Everything else — submissions, interviews, placements — is a derivative of that record. Make the record rich enough that any recruiter on the desk can pick it up cold and pitch the candidate inside ten minutes.

Mandatory fields beyond the basics: source detail (which job board, which referrer, which LinkedIn search), screen date and screen notes, salary history with verifiable data points (last base, last total comp, last bonus), reason for leaving, top three motivators (comp, growth, culture, location, technology, mission), top three dealbreakers, references on file, work authorization with expiration date if applicable, and a timeline of every touchpoint — calls, emails, submissions, interviews. Every field is searchable, every field is reportable.

The most common mistake at small agencies is treating a job order as an email thread. A job order is a contract: it has a client, a fee, terms, a target close, and an aging clock. It needs to live as a record with its own ID, its own status pipeline, and its own dashboard.

Job-order status pipeline: New (just received, not yet worked), Active (recruiters sourcing and submitting), On Hold (client paused), Filled (placement made), Lost (filled by another agency or client withdrew), and Expired (aged out without a fill). Most agencies track 'time-in-stage' on each job order — a job order sitting at New for more than 48 hours is a leading indicator of poor desk discipline.

For every active job order, the system should display: open date, days open, fee amount and structure, hiring manager, candidates submitted, candidates interviewing, candidates with offers, and the named recruiter on point. A weekly job-order review meeting where the leadership team works through every active order in this format takes 30 minutes for a 25-order desk and prevents nearly every stalled-order problem before it becomes a lost-order problem.

A submission is a candidate-to-job-order pairing with a date, a submitted-rate, and a status. Submissions are the operational heart of the agency — they are how recruiters demonstrate work, how managers measure activity, and how the system computes contingent fees.

Every submission record needs: candidate ID, job order ID, submission date, submitting recruiter, submitted salary or rate, submission method (direct email, client portal, ATS), client recruiter or hiring manager who received it, status (under review, interview scheduled, declined, hired), and a feedback note if the client declined. Submission velocity (submissions per recruiter per week) and submission-to-interview conversion rate are the two metrics that predict placement volume four weeks out — if these are off, billings will be off, and you will see it before anyone misses a number.

Interview coordination is where placements quietly die. The candidate gets a confusing calendar invite, the hiring manager reschedules without telling the recruiter, the candidate's current employer schedules an offsite the same day. Treat interview coordination as a workflow with named steps and SLAs.

For every interview round, the record should capture: round number, type (phone, video, technical, panel, final), date and time with timezone, interviewers (names and titles), location or video link, prep materials sent to candidate, debrief notes from candidate, debrief notes from client, and outcome. Send the candidate a prep packet the day before with the interviewers' names, LinkedIn profiles, the role description, and three prepared questions to ask. Debrief the candidate within two hours after the interview and the client within one business day. Skipping either debrief is the failure mode that produces 'we went with another candidate' emails with no actionable feedback.

The placement is the billable event. It needs its own record — not a status update on the candidate or the job order, but a discrete record that the finance team can audit and the recruiters can see in their commission view.

A placement record contains: candidate ID, job order ID, client company, accepted offer amount (base + variable), start date, employment type, fee percentage or flat amount, fee currency, invoice date, invoice number, payment terms, billing recruiter, splitting recruiter (if any), split percentage, gross fee, recruiter commission, manager override, and house margin. The system should auto-generate the invoice from the placement record on the start date (or per client terms — some clients invoice on offer-acceptance), and feed the placement into the recruiter's commission ledger.

Commission structures vary, but the common patterns are: flat percentage of gross fee per recruiter (typical 25-50%), tiered percentage based on monthly billings, and split commissions for accounts where one recruiter owns the client and another sources the candidate. Whatever the structure, the system must calculate it deterministically — every recruiter should be able to see at any moment what their year-to-date commission is, what is at risk in the guarantee window, and what is committed and paid.

Most permanent placement contracts include a guarantee period — typically 30, 60, or 90 days — during which if the placed candidate leaves (or is terminated), the agency owes a free replacement, a prorated refund, or a full refund. The guarantee period is when placements stop being billable income and start being potential refunds. The system has to track this explicitly.

For every placement, store the guarantee end date as a calculated field from start date plus contract terms. Surface guarantee-window placements on the recruiter's and the agency's dashboard. Track candidate check-ins at 30, 60, and 90 days — a placed candidate is the agency's strongest source of referrals if treated like a relationship, and the earliest detection signal for an at-risk placement if treated like a checkbox. Reserve commission accruals against placements still in guarantee window — paying full commission on day-one and clawing it back on day-45 is one of the fastest ways to destroy recruiter trust in the commission system.

Deelo is built for agencies that want one platform for the BD pipeline, the candidate database, the job-order desk, and the placement and commission ledger — without stitching together a separate ATS, CRM, and accounting tool. The CRM provides the candidate, contact, and company records with custom fields for the schema decisions described above. Practice/Matters models job orders and placements as first-class records with their own status pipelines and aging clocks. The Automation app handles guarantee-window check-ins, submission-velocity alerts, and stalled job-order warnings without a separate workflow tool. Docs and ESign handle engagement letters, fee agreements, and offer paperwork. The Invoicing app turns placements into invoices automatically. The client portal lets hiring managers see live candidate status without logging into a third-party tool.

For a 5-recruiter agency replacing a fragmented stack of spreadsheets, an entry-level ATS, a separate CRM, and a Zapier subscription tying them together, the consolidation usually drops monthly tool spend by 40-60% while eliminating the data-sync errors that produce miscalculated commissions. For agencies above 15 recruiters, the value is less about tool consolidation and more about having a single system of record that the leadership team can run dashboards against without joining four different exports.