Lever's Opportunity model must split into Greenhouse Candidate + Application records. Scorecards can't be created via API — import as notes. Budget for rate limits on both sides and download Lever resume URLs immediately.
There is no native migration path from Lever to Greenhouse. Lever is opportunity-centric — a single Contact has multiple Opportunities representing candidacies. Greenhouse separates this into distinct Candidate and Application records where Applications link Candidates to Jobs with structured scorecards. The fundamental gap is that Greenhouse's Harvest API does not expose a POST endpoint for creating scorecards. Interview feedback from Lever cannot be migrated as native Greenhouse scorecards — only as structured notes on the activity feed. Every migration requires handling Lever's temporary resume URLs, the On-Behalf-Of header on all Greenhouse writes, strict rate limits on both sides, and the structural translation from a relationship-first model to a process-enforcement model.
Read this first
Pair-specific gotchas that catch teams out. Each one has cost somebody a weekend.
Do not collapse Lever's contact and opportunity into a single flat row
Keep two immutable source keys throughout the project: one for the person (contact) and one for the candidacy (opportunityId).
Scorecard limitation
Greenhouse's Harvest API provides GET endpoints for scorecards but does not expose a POST endpoint to create them. Interview feedback from Lever cannot be migrated as native Greenhouse scorecards. The standard workaround is to import feedback as structured notes on the candidate's activity feed, preserving the interviewer name, rating, and per-attribute feedback as formatted text. (developers.greenhouse.io)
Greenhouse's bulk import supports custom candidate and application fields as additional
Greenhouse's bulk import supports custom candidate and application fields as additional columns, but the fields must already exist in Greenhouse and the values must pass validation. Historical imports can trigger GDPR or CCPA communications depending on your tenant configuration. (support.greenhouse.io)
Lever resume URLs are temporary
Download resume and file content immediately during extraction. Do not store URLs for later retrieval — they will expire. Lever can also return a 422 when a file could not be processed correctly. (hire.lever.co)
Harvest v1/v2 API deprecation deadline
August 31, 2026. If you're building a migration pipeline today, build against v3 (OAuth). The auth model is fundamentally different — v3 requires OAuth token lifecycle management instead of static API keys.
Always run a test migration first
Use Greenhouse's sandbox environment. Object IDs differ between sandbox and production — don't hardcode IDs from test runs. Run at least two dry runs: one happy-path job and one ugly job with feedback, files, duplicates, and custom fields.
The runbook
Work top to bottom. Tick steps as you go — your progress is saved in this browser.
01 Discovery Scope candidates, pipelines and the compliance obligations that come with them.
Objective Agreed scope across candidates, applications, jobs and interview history, with legal signed up on retention.
Keep these open
-
Inventory every object in Lever
Count candidates, applications (a candidate can have many), jobs and requisitions, interviews, scorecards, offers, and resume files. Applications and scorecards usually outnumber candidates several times over, and resume files dominate storage.
Data Profiler Get real record counts instead of estimating from memory -
Settle retention and consent with legal
Candidate data is heavily regulated: GDPR right-to-erasure, EEOC/OFCCP record-keeping, and per-region retention windows that conflict with each other. Decide what may be migrated at all before you scope anything else — this frequently shrinks scope substantially.
Migrating candidate records whose consent has lapsed or whose retention window has expired creates a new compliance breach in the target system.
-
Map the hiring pipeline and agree the target stages
Document every job's pipeline, stage, and rejection reason, then agree the Greenhouse model with talent leadership. Stage definitions drive every funnel metric you report, so changing them silently rewrites your hiring analytics.
-
Catalogue integrations and the job-board estate
List job boards, careers-site integration, HRIS, background check, assessment platforms, calendar and email. The careers site and job boards are customer-facing, so their cutover needs its own plan and its own testing.
-
Build the business case and choose the window
Model licence delta, effort and recruiter productivity dip. Time the window against your hiring cycle: a migration during peak graduate recruitment or a hiring surge will fail on people, not technology.
Vendor Evaluator Score Greenhouse against alternatives on weighted criteria
Lever → Greenhouse specifics
- Structured hiring at scale
- Greenhouse enforces consistent, auditable interview processes across departments and geographies. Its scorecard system, approval workflows, and configurable interview plans are more rigid than Lever's — by design.
- Integration ecosystem
- Greenhouse integrates with over 1,000 job boards and has a broader partner network for background checks, assessments, and HRIS connectors. Companies scaling past ~200 employees often find Greenhouse's ecosystem covers more of their tool stack.
- Compliance and DEI
- Greenhouse offers structured EEOC data collection, anonymized interview scoring, and built-in diversity analytics. For companies in regulated industries or with formal DEI mandates, these features reduce the compliance engineering burden.
- Small team, basic data
- CSV import gets you running in hours. Accept some manual cleanup.
- Enterprise, zero downtime
- Managed service — the engineering cost of handling all edge cases in-house rarely justifies the investment.
Don't move on until
- Counts confirmed for candidates, applications, jobs and offers
- Retention and consent obligations confirmed with legal
- Hiring-stage model agreed with talent leadership
02 Data Audit Audit candidate data with compliance sitting next to you.
Objective Profiled exports with duplicates, expired records and resume files all quantified and triaged.
Keep these open
-
Export and profile candidates, applications and jobs
Profile each object separately and reconcile against API counts. Watch the candidate-to-application ratio: a mismatch usually means applications have been silently truncated by pagination.
Data Profiler Profile the Lever export for nulls, outliers and type drift -
Quantify duplicate candidates and agree survivorship
The same person applies repeatedly over years with different emails and name spellings. Measure the duplicate rate and agree survivorship rules — which record wins and what happens to the application history attached to the losers.
Merging candidates without agreed survivorship rules destroys application and interview history that you may be legally required to retain.
Data Cleaner Strip empty rows, stray whitespace and dead columns -
Identify records outside their retention window
Flag candidates whose consent has expired, who have exercised erasure, or who fall outside regional retention. Exclude them from scope and document the exclusion — you need to show the decision was deliberate.
PII & Compliance Scanner Find regulated fields before they land in a new system -
Inventory resume files and attachments
Count files, total volume and MIME types, and check every attachment still resolves to a live URL. Expiring signed download URLs are the classic reason a resume migration completes with a large fraction of empty files.
Resume download URLs are often short-lived signed links. Fetch files close to load time or they will 404 mid-migration.
-
Verify relationship integrity
Confirm every application links to a live candidate and a live job, and every scorecard to a real interview. Orphaned applications produce a funnel report that does not tie to anything.
-
Clean, normalise and produce masked test data
Normalise emails, phone formats and locations, standardise timestamps to UTC, and generate a masked dataset for the sandbox. Real candidate data in a sandbox is a compliance breach in most jurisdictions.
PII Masker Generate a safe copy for sandbox and vendor testing
Lever → Greenhouse specifics
- Authentication
- Basic Auth (API key as username, blank password) or OAuth 2.0
- Rate limit
- 10 requests/second steady state, bursts up to 20 req/sec per API key. Application POST requests are limited to 2 req/sec. (hire.lever.co)
- Pagination
- Offset-token based. Each response includes hasNext and next fields. Page size configurable from 1–100.
Don't move on until
- Duplicate candidate rate quantified with a merge policy agreed
- Records outside retention identified and excluded
- Resume and attachment inventory complete with total volume
03 Field Mapping Map the candidate-application-job triangle before anything else.
Objective A signed mapping covering objects, stages, rejection reasons, scorecards and EEO fields.
Keep these open
-
Map the candidate, application and job model
ATS platforms differ on whether a person or an application is the primary record. Establish this first: getting it wrong means one candidate becomes five, or five applications collapse into one, and the entire field map has to be redone.
Candidate-centric and application-centric models are not interchangeable. Confirm which Greenhouse uses before mapping any field.
Schema Mapper Opens pre-loaded with the Lever → Greenhouse field pair -
Map pipeline stages and rejection reasons exhaustively
Enumerate every stage and rejection reason across all jobs, including retired values on historical applications, and map each explicitly. Unmapped rejection reasons are both a reporting gap and, in regulated hiring, a compliance one.
-
Decide EEO and diversity data handling
These fields are separately regulated and often legally required to be stored apart from the candidate record. Confirm with legal whether they migrate at all, and how Greenhouse isolates them.
EEO data usually cannot be migrated into ordinary custom fields without breaching the segregation rules that govern it.
-
Map interviews, scorecards and feedback
Structured scorecards rarely have a native equivalent. Decide whether to reconstruct them, flatten them into notes, or keep them only in the archive — and be explicit that flattening loses the ability to report on them.
JSON to CSV Converter Flatten nested API responses into a reviewable sheet -
Plan resume and file migration
Confirm size limits, MIME support and whether Greenhouse re-parses resumes on upload. Re-parsing can overwrite carefully curated candidate fields with worse machine-extracted values, so test it deliberately.
-
Set load order and freeze the spec
Users, then jobs, then candidates, then applications, then interviews and scorecards, then files. Keep source IDs in custom fields, then version and sign off the spec.
Lever → Greenhouse specifics
- Name splitting
- Lever stores name as a single string. Greenhouse requires first_name and last_name.
- Email restructuring
- Lever's emails [] is a flat string array. Greenhouse expects {value, type} objects.
- Timestamp conversion
- Lever uses Unix millisecond timestamps. Greenhouse uses ISO 8601.
- Stage mapping
- Build a lookup table mapping Lever stage IDs → Greenhouse stage IDs.
- Source mapping
- Match Lever sources to Greenhouse source IDs. Create missing sources first.
Don't move on until
- Candidate/application/job model mapped and reviewed
- Every stage and rejection reason explicitly mapped
- EEO and diversity field handling agreed with legal
04 Test Migration Pilot whole candidate journeys, not isolated records.
Objective A sandbox pilot where candidate journeys, funnel metrics and resume files all verify.
Keep these open
-
Configure the Greenhouse sandbox with the agreed pipelines
Create jobs, pipeline stages, scorecard templates, user roles and custom fields first. Loading applications before the stages exist puts every candidate in a default stage and invalidates the pilot.
-
Select complete candidate journeys as the pilot slice
Take 100-200 candidates with all their applications, interviews, scorecards and files — including repeat applicants, hires, rejections at every stage, and candidates on multiple jobs. Repeat applicants are where the model mapping actually gets tested.
-
Run the load in dependency order with logging
Jobs, candidates, applications, interviews, then files, logging each request against its source ID. Track file uploads separately: they fail for different reasons and at different rates than record writes.
-
Verify journeys and funnel metrics
Confirm each candidate sits at the right stage on the right job with their history intact, and that per-stage funnel counts match Lever for the pilot jobs. Funnel mismatches point straight back to stage mapping.
Migration Validation Tool Diff the pilot batch against source before scaling up -
Open every pilot resume and check re-parsing
Actually open the files rather than trusting the upload count, and check whether re-parsing has overwritten any candidate fields. A resume that uploaded as a zero-byte file still counts as a success in most logs.
-
Put recruiters in front of the pilot data
Have recruiters work their own pilot requisitions end to end. They immediately spot missing feedback, wrong stages and unreadable history that reconciliation cannot see.
Don't move on until
- Candidate-application-job relationships intact for the pilot
- Funnel counts per stage match source for pilot jobs
- Resume files open correctly for every pilot candidate
05 Cutover Switch recruiting without dropping a live candidate.
Objective All in-scope recruiting data live in Greenhouse, careers site and boards repointed, recruiters working.
Keep these open
-
Pre-load historical candidates and closed jobs
Load closed requisitions, rejected candidates and archived applications while Lever stays live. Only active pipeline and the final delta need to move in the window.
-
Publish the runbook including the careers-site switch
A timed sequence with owners and abort criteria, treating the careers site and job boards as first-class steps. They are candidate-facing, so a failure there is publicly visible in a way a data defect is not.
-
Freeze Lever and take the final delta
Stop new applications and let recruiters finish in-flight actions, then export everything changed since the pre-load. Coordinate with anyone actively interviewing so feedback is not entered into the old system mid-freeze.
Interview feedback entered in the old ATS during the freeze is lost, and it is the data recruiters most immediately notice missing.
-
Load active pipeline and reconcile stages
Load active candidates and applications, then verify every active candidate is at the correct stage on the correct job before repointing anything. Active-stage accuracy is what recruiters check first on day one.
Migration Validation Tool Confirm the final delta landed before you reopen -
Repoint careers site, job boards and integrations
Switch the careers-site integration, repost or migrate live job ads, and repoint HRIS, background check, assessment and calendar integrations. Then submit a real test application through the careers site and every major board.
Job ads left posted against the old ATS keep collecting applications that never reach the new system.
-
Go/no-go and switch recruiters over
Call the decision against the exit criteria, then move recruiters with support on hand for the first day. Keep Lever read-only — candidate records have retention obligations that outlast the migration.
Don't move on until
- Historical load complete and reconciled before the freeze
- Careers site and job boards posting into Greenhouse and verified
- Active candidates confirmed at the correct stage
06 Validation Prove the funnel, the files and the compliance position.
Objective Reconciled recruiting data, funnel parity with baseline, and a defensible compliance record.
Keep these open
-
Reconcile every object including files
Compare counts for candidates, applications, jobs, interviews, scorecards and files, with field-level spot checks on a sample. Count files separately — they are the object most likely to be quietly short.
Migration Validation Tool Reconcile Lever and Greenhouse record-for-record -
Tie funnel and time-to-hire reporting to baseline
Rebuild funnel conversion, time-to-hire, source effectiveness and offer-acceptance reporting and compare to pre-migration figures. Variances trace back to stage mapping and to how application timestamps were handled.
Time-to-hire depends on stage-transition timestamps. If those were approximated, the metric will differ even with identical records.
-
Verify file integrity at scale
Sample-open resumes across the whole load and compare file sizes against source. Zero-byte and truncated files are common and never surface in an upload success count.
-
Confirm retention, consent and EEO configuration
Verify retention rules, consent state and EEO segregation are correctly configured in Greenhouse, and that excluded records genuinely did not migrate. File this as your compliance evidence.
PII & Compliance Scanner Produce the compliance evidence your auditor will ask for -
Test workflow, notifications and candidate-facing paths
Fire every stage automation, interview scheduling flow, rejection template and offer approval, and submit a live application through the careers site. Candidate-facing emails going out wrong is a brand problem, not just a bug.
-
Sign off and schedule decommission
Get written acceptance against the Discovery criteria, retain Lever read-only for the period your retention policy requires, take a final archive export, and diarise cancellation.
Don't move on until
- Reconciliation complete across candidates, applications and files
- Funnel and time-to-hire reporting tie to baseline
- Retention and EEO configuration verified, acceptance signed
Field mapping reference
The field-by-field mapping for each object. Use this as the starting point for your mapping spec.
Object Equivalent
| Lever field | Greenhouse field | Notes |
|---|---|---|
| Contact | Candidate | 1:1 mapping. Use Lever's contact field as the dedup key. |
| Opportunity | Application | Each Lever Opportunity maps to one Greenhouse Application on a Job. |
| Posting | Job + Job Post | Lever combines the internal job config and public posting. Greenhouse separates them. |
| Feedback (per Opportunity) | Scorecard | Read-only in Greenhouse API. Cannot create scorecards via API — must import as notes. |
| Notes | Activity Feed (Notes) | Map to POST /candidates/{id}/activity_feed/notes. |
| Interview Panels | Scheduled Interviews | Greenhouse requires an application_id and valid interviewers. |
| Offers | Offers | Field-level mapping required. Offer custom fields only available on Enterprise tier. |
| Requisition | Job Opening / Requisition | Lever requisitions map to Greenhouse openings + requisition IDs on jobs. |
| Tags | Candidate Tags | Direct mapping. |
| Archive Reasons | Rejection Reasons | No 1:1 mapping. Build a lookup table. |
| Sources | Sources | Create matching sources in Greenhouse before import. |
| Resume Files | Attachments | Lever resume download URLs are temporary. Download immediately during extraction. |
Risk matrix
Per-object risk for this pair. Plan extra validation around anything marked high.
| Object | Risk | Notes |
|---|---|---|
| Scorecards/Feedback | high | No API creation endpoint — must import as structured notes, losing native queryability |
| Attachments/Resumes | high | Lever URLs are temporary; must download immediately and re-upload as base64 |
| Interview Scheduling History | high | Calendar events and panel configurations have no import path in Greenhouse |
| EEOC/Demographic Data | high | Lever stores at opportunity level; Greenhouse has different privacy model requiring legal review |
| Confidential Postings | medium | Lever requires special API key permission at creation time; cannot be added retroactively |
| Custom Fields | medium | Application-level custom fields require Greenhouse Enterprise tier; type mismatches common |
| Pipeline Stage History | medium | Stage names rarely map 1:1; requires semantic mapping via lookup tables |
| Source Attribution | low | Direct mapping with pre-creation of matching sources in Greenhouse |
| Tags | low | Direct copy from Lever opportunity tags to Greenhouse candidate tags |
| Notes/Activity Feed | low | Maps cleanly to Greenhouse candidate activity feed notes endpoint |
The hard parts
What makes this specific migration difficult, beyond the mechanics.
No Scorecard Creation API
Greenhouse exposes GET endpoints for scorecards but no POST endpoint. Lever feedback must be imported as structured notes, losing native scorecard queryability.
Multi-Opportunity Splitting
A single Lever Contact with multiple Opportunities must become one Greenhouse Candidate with multiple Applications. Incorrect handling creates duplicate profiles that cannot be bulk-merged.
Temporary Resume URLs
Lever hosts attachments via temporary download URLs. Files must be downloaded during extraction and re-uploaded to Greenhouse as base64 or staged URLs.
On-Behalf-Of Requirement
Every Greenhouse write operation requires a valid On-Behalf-Of header with an active user ID. Deactivated employees referenced as authors require workarounds.
Custom Field Type Mismatches
Lever custom fields live at the Opportunity level while Greenhouse splits across candidate, application, job, and offer levels. Application-level fields require Enterprise tier.
Tools used in this playbook
All free, all run entirely in your browser — nothing is uploaded.
FAQ
How does Lever's data model map to Greenhouse?
A Lever Contact maps to a Greenhouse Candidate (1:1). Each Lever Opportunity maps to one Greenhouse Application tied to a specific Job. A single Contact with multiple Opportunities becomes one Candidate with multiple Applications. Keep both source IDs as immutable external keys for reconciliation.
Can you migrate interview scorecards from Lever to Greenhouse?
Not as native scorecards. Greenhouse's Harvest API does not expose a POST endpoint for creating scorecards — only GET endpoints for reading them. The standard workaround is to import Lever feedback as structured notes on the candidate's activity feed, preserving interviewer name, overall recommendation, and per-attribute ratings.
What is the Greenhouse Harvest API rate limit?
For Harvest v1/v2, the rate limit is typically 50 requests per 10-second window. Exceeding this returns an HTTP 429. Official docs say 429 responses include Retry-After and X-RateLimit-Reset headers, but X-RateLimit-Limit and X-RateLimit-Remaining may be absent. Harvest v3 uses a fixed 30-second window. Note that v1/v2 will be deprecated on August 31, 2026.
Can Zapier or Make handle a Lever to Greenhouse migration?
They can help with low-volume delta syncs during a transition period, but they are not viable for bulk historical migration. Zapier's Greenhouse triggers use polling rather than instant webhooks, neither tool natively handles the On-Behalf-Of header required for Greenhouse writes, and they lack the state management needed for large backfills.
How long does a Lever to Greenhouse migration take?
It depends on dataset size and complexity. A small team (<500 candidates) using CSV import can finish in hours. API-based migrations of 10,000–50,000 candidates with feedback, attachments, and custom fields typically take 40–200 hours of engineering time for DIY, or 2–5 days with a managed migration service.