Migration Checklist: Moving to Recruit CRM

Phase 0

Decide & Plan

Lock scope, ownership, and constraints before any export job runs.

• Document the in-scope Recruit CRM modules. Decide which of Candidates, Contacts, Companies, Jobs, Deals, Tasks, Meetings, Notes, Call Logs, Hotlists, Placements, Invoices, and Timesheets are part of this project. Record each as in or out of scope.

  Risk if skipped: Mid-project scope additions force re-running mapping, sandbox tests, and Hiring Pipeline design from the start.

• Define historical-data scope per object. Decide how many years of candidates, contacts, jobs, notes, call logs, meetings, and email history move across. Record cut-off dates per object so the export job is bounded.

  Risk if skipped: Without an explicit cut-off, the export expands into a multi-year dump that no one budgeted for transforming.

• Decide whether candidate Hiring Stage history is in scope. Recruit CRM does not preserve historical Hiring Stage transitions through bulk import. Choose between replaying transitions in the destination to rebuild the trail, dumping the trail into a long-text custom field, or accepting current stage only.

  Risk if skipped: Discovering this constraint mid-cutover forces an emergency decision that loses years of placement-funnel analytics.

• Confirm base-plan record and custom-field caps. The base plan caps candidate records at 10,000 and custom fields at 15 across modules. Confirm whether the target tier raises these caps or whether scope must be trimmed to fit.

  Risk if skipped: Loading past the record or field cap halts the import partway and forces unplanned plan changes during the cutover window.

• Assign a single accountable migration owner (DRI). Name one person responsible for go/no-go on cutover. Document an escalation chain that includes Recruiting Operations, Sales (for Deals), IT, and Finance representatives.
• Capture GDPR, residency, and retention obligations. Recruit CRM hosts data on AWS with AES-256 encryption at rest. Document the Data Processing Agreement, transfer-impact assessment, candidate-consent provenance, and any retention rules that change on cutover.

  Risk if skipped: Legal blocks the import on go-live day when candidate-consent provenance or DPA review was not completed earlier.

• Set the cutover date, source freeze window, and parallel-run window. Choose a date that avoids active client placements, candidate offer-stage activity, and invoicing periods. Define the freeze window during which the source ATS goes read-only.

  Risk if skipped: A cutover during an active placement push forces a partial cutover and a hand-reconciled candidate pipeline.

• Document the migration success criteria in writing. List the row counts, candidate-pipeline reconciliations, and recruiter-acceptance checks that must pass before the project is closed. Include the source-system decommission date.
• List explicitly out-of-scope items. Examples: legacy email sequences, archived candidate pools, source-side automations, sourcing-extension caches. Capture them so they do not silently re-enter scope mid-project.

Phase 1

Pre-Migration Prep

Source-system cleansing and destination-tenant configuration before any export.

Phase 2

Source Export

Pull every object you will need out of the source system in a transformable shape.

• Export the master candidate list to CSV with all standard and custom fields. Include first name, last name, primary email, phone, current employer, current title, location, salary expectations, Source, current Hiring Stage, owner, and every custom field from the Phase 1 audit.
• Export the master contact list to CSV. Include first name, last name, email, phone, title, linked company name, owner, and every contact-side custom field. Contacts are the hiring-manager and client-side records.
• Export the company list to CSV. Include company name, domain, industry, address, owner, and every company-side custom field. Capture Parent/Child relationships so they can be rebuilt post-load.

  Risk if skipped: Without Parent/Child mapping data, multi-entity client groups collapse into a flat company list and break account-level reporting.

• Export the job list to CSV with pipeline assignment. Include job title, linked company, hiring manager, status, open date, close date, salary range, location, assigned recruiter, Hiring Pipeline name, and every job-side custom field.
• Export deal records to CSV. Include deal name, linked company, deal stage, value, currency, expected close date, owner, and every deal-side custom field.
• Export Hiring Stage history per candidate per job. For every candidate that touched a job, export the full sequence of stage transitions with stage name, timestamp, and the user who made the move. This is the input to the Phase 3 history-rebuild decision.

  Risk if skipped: Without the transition log, source-of-hire and time-to-fill reports cannot be reconstructed in the destination.

• Export placement records. Per placement: candidate, job, start date, fee structure, commission split, and invoice status. Recruit CRM uses Placements as the bridge between Jobs and revenue tracking.
• Export candidate-to-job association records. Per candidate, export every job they are linked to with current stage on each. This is the junction data that drives the candidate's Jobs tab in Recruit CRM.

  Risk if skipped: Loading candidates and jobs separately without the association file leaves the destination Kanban boards empty.

• Export notes, meetings, tasks, and call logs as separate files. Each object has a different schema. Include the parent record id, timestamp, owner, and content. Tag each row with whether it relates to a candidate, contact, company, job, or deal.
• Export email correspondence as a per-record file. Include sender, recipients, subject, body, sent date, and the linked candidate or contact. Decide whether to load as Notes attached to the timeline or to archive separately.

  Risk if skipped: Importing thousands of emails as Notes can blow past per-record activity limits and slow timeline rendering for the rest of the system's life.

• Export Hotlist memberships per candidate. For each Hotlist, export the candidate ids on the list with the date they were added. Hotlists are lightweight collections that drive ongoing-role and talent-pool workflows.
• Export candidate resume files with metadata. Pull resume binaries (PDF, DOCX) and a metadata file with candidate id, original filename, upload date, and whether the resume is the latest version.

  Risk if skipped: Without metadata you have a folder of resumes no one can re-attach to the right candidate.

• Export candidate and company attachments beyond resumes. Pull signed contracts, offer letters, right-to-work documents, and any portfolio uploads. Capture the linked record id, original filename, and document category.
• Export the user list with roles and team assignments. Capture which users were recruiters, sourcers, account managers, or admins in the source so Phase 1 user-creation stays grounded in reality.
• Export pipeline and workflow definitions. Pull every Hiring Pipeline configuration, every Deal Stage list, and any workflow-automation rules. Configurations do not transfer; they will be rebuilt.
• Capture screenshots of source dashboards and reports. Time-to-fill, source-of-hire, placement-revenue, and pipeline-funnel reports will be rebuilt by hand. Screenshot the definitions and filter criteria.
• Capture screenshots and configuration exports of source automations. Email sequences, candidate-status auto-updates, SMS reminders, and integration hooks will be rebuilt by hand. Document the trigger conditions and the actions taken.
• Record the exact timestamp of each export. Log the date and time per export file. The delta between this timestamp and cutover is what you re-export on cutover day.

Phase 3

Transform & Map

Turn raw source files into Recruit CRM-shaped imports.

Phase 4

Sandbox Test Migration

Catch every problem before the Bulk Importer runs against production.

• Set up a sandbox or trial tenant configured identically to production. Mirror the Phase 1 setup: same users, same Hiring Pipelines, same custom fields, same picklist values. Keep both tenants in lockstep so sandbox results extrapolate.

  Risk if skipped: A sandbox that drifts from production hides bugs that only fire under the production schema.

• Load a 1–5% candidate sample through the Bulk Importer. Pick a sample that covers active candidates, placed candidates, rejected candidates, and candidates with non-Latin characters in their name so edge cases surface early.

  Risk if skipped: A sample of only active North American candidates hides the bugs that hit the harder cases in production.

• Walk the recruiter who will run the import through the Bulk Importer UI. Confirm the operator understands the Merge Duplicates match-key picker and the Override Data toggle before they touch a real file.

  Risk if skipped: An accidental Override Data import overwrites manually maintained destination records with stale source values.

• Validate 10 random sandbox candidates field by field. Compare each value side-by-side against the source. Note date, currency, Source-field, and custom-field mismatches.
• Test edge cases: names with non-ASCII characters, multi-line addresses, leading zeros in IDs. Confirm the importer round-trips special characters and preserves leading zeros on text-typed numeric fields.
• Test resume-file attachment as documents without parser auto-run. Upload sample resumes against the sandbox candidates and confirm the AI Resume Parser does not overwrite candidate fields you populated from the source CSV.

  Risk if skipped: If the parser fires by default, plan a per-candidate field-restore pass before production.

• Test loading a sample job with a pre-existing Hiring Pipeline assigned. Confirm the job lands with the correct Hiring Pipeline and that every stage is visible on the Kanban board.
• Test the Hiring Stage history rebuild pattern end-to-end. If using the replay pattern, throttle the loader and confirm transitions land in the correct order. If using the paragraph-field pattern, confirm the long-text field renders the transition log correctly.

  Risk if skipped: Effective-date order issues only show up in the UI; queries return rows in id order, which masks the problem until UAT.

• Test the candidate-to-job association load. Confirm the candidate appears on the job's Kanban at the right stage and that the candidate's Jobs tab lists the job correctly.
• Test Hotlist membership loading. Add a sample of candidates to a sandbox Hotlist and confirm the membership renders on the Hotlist page.
• Test Parent/Child Company linkage on a small set. Link two sandbox companies as parent/child and confirm both pages display the relationship correctly.
• Test interview meeting and Feedback Form load. Load a sample interview with attached scorecard responses and confirm both render on the candidate's profile and the job's panel view.
• Run a full-volume dry run within the sandbox throughput budget. Load the entire candidate and job set in the sandbox and time the run. Multiply by a safety factor to forecast the cutover-window duration.

  Risk if skipped: A cutover window that turns out to be too short forces a partial go-live with candidates split between systems.

• Capture every error code and row id from the Bulk Importer logs. Categorize errors as recoverable (fix the transform) vs. non-recoverable (drop the row, document the reason).
• Have a domain expert review 25 sandbox candidates and 10 jobs. Pick a senior recruiter who is not on the migration team. Their job: spot anything that 'looks wrong' to a daily user.
• Get written sign-off on sandbox results before cutover scheduling. The DRI signs off on row counts, stage-history reconciliation, resume-attachment integrity, and the sandbox UAT report.

  Risk if skipped: Skipping sign-off lets the cutover proceed with known sandbox defects still unresolved.

Phase 5

Production Cutover

The tightly-sequenced execution window where the migration actually happens.

• Confirm the source-system freeze is in effect. Verify there are no candidate, job, or deal changes in the source in the hour before export.

  Risk if skipped: Late source-side edits create a delta that the cutover plan has no time to reconcile.

• Take a final full export of source data. Re-run every export from Phase 2 at the freeze time so the import file reflects the exact state at cutover.
• Confirm Chrome Extension, AI Voice, and sequences are still paused. Re-verify the paused state on every automation that could fire on imported records.

  Risk if skipped: An unpaused sequence fires during the bulk load and sends weeks-old outreach to candidates the moment their record lands.

• Load users and team assignments first. If users were created in Phase 1, only refresh team and role assignments now. If not, create users before any record load so ownership can attach correctly.
• Load Companies first via the Bulk Importer. Use email-domain or normalized-name as the merge key. Stamp the source id into the Legacy ID custom field on every row.
• Load Contacts second, linked to Companies by legacy id. Match Contacts to Companies via the Legacy Company ID field. Confirm every contact lands with its linked company name visible.
• Apply Parent/Child Company linkages in a third pass. Once all companies exist, run the Parent/Child link action per pair via the company page.
• Load Candidates next via the Bulk Importer. Use email as the merge key with Merge Duplicates on and Override Data set per the Phase 3 decision. Stamp Legacy ID on every row.
• Load Jobs next, linked to Companies and Contacts. Match Jobs to their company and hiring-manager contact via legacy ids. Confirm every job lands with its Hiring Pipeline assigned.

  Risk if skipped: A job that loads without a pipeline strands every linked candidate outside the Kanban board.

• Load candidate-to-job associations with current Hiring Stage. Write each association in. Throttle the loader to fit within the platform's documented throughput budget.
• Replay Hiring Stage history in chronological order, throttled. If using the replay pattern, walk each candidate-job pair's transition list ascending by timestamp and write each transition.

  Risk if skipped: Out-of-order replay loads transitions backwards and the final stage no longer reflects the candidate's true current state.

• Load Deals linked to Companies and Contacts. Match Deals to their company and primary contact via legacy ids. Confirm every deal lands with the correct deal stage.
• Load Placements linked to Candidates and Jobs. Placements bridge revenue tracking. Confirm fee structure and commission split land intact.
• Attach resume binaries as documents with parser auto-run disabled. Upload one resume per candidate. If the AI Resume Parser is configured to auto-run on upload, batch-run a re-stamp pass after the load to restore overwritten fields.

  Risk if skipped: Letting the parser run during cutover overwrites manually verified candidate fields with raw resume text.

• Attach non-resume documents in a separate batch. Load offer letters, signed contracts, right-to-work documents, and portfolio uploads as document files, tagged by category.
• Load Hotlist memberships per candidate. Add candidate ids to each Hotlist. Confirm membership counts match the export.
• Load interview meetings and Feedback Form responses. Create each meeting linked to its candidate, job, and panel. Attach Feedback Form scorecards per the Phase 3 decision.
• Load notes, calls, and tasks linked to their parent records. Create each note, call log, and task with the candidate, contact, company, job, or deal id. Time-throttle the loader.
• Capture row counts, error logs, and job ids for every load step. Save the Bulk Importer history and the load job ids in the migration runbook for audit and rollback evidence.

  Risk if skipped: Without per-step evidence a partial failure has no recovery point and the team starts the load over from row one.

• Re-enable Chrome Extension, AI Voice, and sequences. Resume each automation only after every load step is reconciled. Confirm sequences do not back-fire on imported records by adjusting their trigger conditions if needed.

Phase 6

Validate

Prove the migration was correct before any user logs in.

• Reconcile total candidate count between source and Recruit CRM. Pull the candidate count via a custom report. The number must match the Phase 1 baseline minus any agreed exclusions.

  Risk if skipped: An unexplained delta of even a handful of candidates signals a silent drop somewhere in the transform.

• Reconcile total contact, company, job, deal, and placement counts. Per module, compare the destination count to the Phase 1 baseline. Investigate every variance over 0.1%.
• Reconcile active candidate count across every Hiring Pipeline. Sum candidates on each pipeline's Kanban and compare to the source figure. Investigate any pipeline with a delta.

  Risk if skipped: A pipeline-level delta hides a bad Hiring Pipeline assignment that leaves candidates outside any board.

• Reconcile aggregate deal pipeline value. Aggregate deal value by stage and compare to the source figure. Mismatches usually signal a currency-conversion or decimal-separator issue.
• Cross-check 30 random candidates field by field against the source. Pick candidates at random across stages, sources, and recruiters. Walk every field on each profile and confirm parity, including the linked resume.
• Cross-check 20 random jobs against the source. Confirm Hiring Pipeline name, current stage, hiring manager link, company link, and every custom field match.
• Verify candidate-to-job associations on a sample of 25 candidate-job pairs. Open each candidate's Jobs tab and confirm the job appears at the correct Hiring Stage with the correct owner.
• Sample 20 Hiring Stage history rebuilds for ordering and date accuracy. If using the replay pattern, walk each pair's stage-change history and confirm chronological order. If using the paragraph-field pattern, confirm the long-text field renders the full log.
• Verify Parent/Child Company links on every multi-entity client group. Open each parent company and confirm every expected subsidiary appears in the linked list.
• Spot-check 15 candidates' resume attachments for filename and access. Open each resume from the candidate profile, confirm it renders, and confirm the AI Resume Parser did not overwrite the candidate's manually verified fields.

  Risk if skipped: A silent parser overwrite is not visible until a recruiter notices that current title or salary expectations have reset to stale values.

• Spot-check 10 Hotlists for membership counts and date-added accuracy. Per Hotlist, confirm the candidate count matches the export and that the date-added custom field is populated where chronology was preserved.
• Verify Source field distribution against the source-of-hire baseline. Run a Source-field report and confirm the destination distribution matches the source within 1%.

  Risk if skipped: A skewed Source distribution signals a picklist mismatch that silently dropped values across thousands of records.

• Verify interview meetings and Feedback Form scorecards on 10 sampled jobs. Per job, open the panel view and confirm interviews render with the correct candidate, panel members, and scorecard responses.
• Verify Placement records and commission-split accuracy. Cross-check 15 placements between source and destination. Mismatched commission splits will surface as billing disputes within weeks.
• Verify note, call-log, and task association with parent records. Sample 25 activities and confirm each one is attached to the correct candidate, contact, company, job, or deal.
• Trigger a Full Export to capture a post-cutover snapshot. Run Admin → Account → Full Export to produce CSV files plus all attached binaries. This becomes the rollback baseline. Full Export is available once every 90 days.

  Risk if skipped: Without a Full Export captured immediately after cutover, a later data-corruption event has no clean restore point until the next 90-day cycle.

• Run a GDPR and SOC 2 evidence collection pass. Capture the import job ids, who ran them, and when. File alongside the DPA and transfer-impact assessment captured in Phase 0.
• Have 3–5 recruiters complete their day-one workflows in Recruit CRM. Cover candidate search, Kanban move, submission to client, Hotlist add, interview schedule, and Feedback Form submit.

  Risk if skipped: Workflow gaps surface only at this layer and are cheaper to fix before go-live than after.

• Document every UAT-surfaced data issue and assign an owner. Triage as block-go-live vs. fix-post-cutover. Block-go-live items go back into the cutover runbook for re-load.
• Get written sign-off from the migration sponsor before Phase 7. Sponsor signs off on row-count reconciliation, stage-history validation, resume integrity, and the UAT report.

Phase 7

Post-Migration Cleanup

Wrap up so the team can treat Recruit CRM as the system of record.

• Announce go-live to all users with login URL and help contact. Include the Recruit CRM mobile-app download links and a short FAQ for password reset, Kanban basics, and submission workflows.
• Run role-based training sessions. Recruiter Kanban workflows (30 minutes), account manager Deals (20 minutes), admin deep dive (60 minutes). Record each session for new hires.
• Set the source ATS to read-only. Disable writes per the freeze-window plan. Communicate the read-only date to anyone who still references historical data.
• Rebuild source-system reports and dashboards inside Recruit CRM. Time-to-fill, source-of-hire, placement-revenue, and pipeline-funnel reports do not transfer. List the top 10 reports the team relied on and rebuild each via Recruit CRM's reporting UI.

  Risk if skipped: Without rebuilt reports the team falls back to the source system within a week, undermining cutover.

• Rebuild email and SMS sequences from the Phase 2 screenshots. Re-create every sequence in the destination. Re-attach trigger conditions, candidate filters, and step delays. Confirm sequences do not back-fire on imported records.
• Re-deploy the Chrome Extension to every sourcer's browser. Install the Recruit CRM sourcing extension on every active sourcer's Chrome profile and authenticate it against the new tenant. Confirm LinkedIn and Indeed profile capture flows work.
• Reconfigure AI Voice agent profiles and call scripts. Recruit CRM AI Voice agentic outreach configurations do not transfer. Rebuild voice prompts, qualification scripts, and routing rules. Run a paired-test before enabling outbound calls.
• Re-establish downstream integrations. Confirm payroll, calendar sync, video-interview, and BI integrations work against the new tenant. Validate at least one round-trip before disabling the manual export path.
• Run the bulk Update → Owner pass to finalize record attribution. If records were loaded under the Data Migration user, use the Update → Owner bulk action on each module to re-attribute records to their named recruiter.
• Schedule a 30-day post-migration retrospective. Capture what went well, what surprised the team, and which Phase 3 transform rules need updating for any future re-runs.
• Set the source-system decommission date. Typically 90 days after cutover, after the final retention window has closed. Document who signs off on shutdown.
• Archive source-system exports and the migration runbook to long-term storage. Encrypt the archive, store the key separately, and record retention duration per compliance policy.
• Schedule the recurring Full Export cadence. Full Export is available once every 90 days. Calendar the recurring run as the long-term rollback baseline. Data Export (CSV only) runs daily for lighter snapshots.

  Risk if skipped: Without a calendared Full Export cadence, the team forgets the 90-day cycle and a data-corruption event finds no recent restore point.

• Lower the migration service account to read-only or deactivate it. Reduce the Data Migration user's privileges once the audit trail is captured and bulk-update passes are complete.
• Close the migration project formally. Final status report, lessons learned, and a project-closure note distributed to stakeholders.