Greenhouse vs Workable

Greenhouse

Structured interview kits and scorecards, scheduling, reporting and SSO sit in the entry plan and an audit log arrives in the top tier, which is the scorecard discipline I need to make sixty hires defensible. But I found no public information on hiring-manager access scoped to their own vacancies, requisition or offer approvals, or German-language candidate communication, so I cannot tell from the pages whether a decision is attributable to a named person.

Workable

Requisition approvals with budget controls, offer approval flows and granular access rights configured at each stage are exactly what I need before handing a vacancy to a hiring manager, and the interview kits, scorecards and self-scheduled interviews with two-way calendar sync make a decision defensible. Audit trails tracking all changes, with every action logged and attributed whether made by model, agent or person, answer the "who decided" question. I found no public information on bulk actions or on time-to-hire and drop-off reporting per stage.

Greenhouse

The candidate portal with real-time status updates and seamless application is genuinely good candidate experience, and sourcing automation is listed as a paid-tier feature. But I found no public information on multiposting, German boards such as the Bundesagentur für Arbeit, a branded career page, Google for Jobs markup, or mobile application without an account — the distribution half of this criterion rests on nothing I can read.

Workable

A branded careers-page builder with a website jobs widget, distribution to 200+ boards with Indeed Platinum and LinkedIn Preferred partner status, and mobile-optimized applications with auto-fill cover most of the funnel, and candidates can withdraw themselves online. But I found no public information on named German boards such as the Bundesagentur für Arbeit or a major German paid board, on Google for Jobs markup, or on source attribution per applicant — the pieces my German-market hiring actually turns on.

Greenhouse

There is a data-subject request portal, a stated commitment to delete or anonymize when there is no ongoing business need, and customer data queued for deletion ninety days after contract end. I found no public information on a configurable deletion deadline after rejection or talent-pool consent with expiry — the candidate portal works on an opt-out link rather than recorded consent — and the top tier's audit log is not described as recording who opened an applicant file.

Workable

Retention is stated only as holding data as long as necessary, deletion happens on request, and a named Data Protection Officer with free-of-charge GDPR rights gives a documented path for an access or erasure request. I found no public information on an automatic deletion deadline after rejection, on talent-pool consent with expiry, or on a log of who viewed an applicant file — the GDPR automation tools are named without detail. For the liability I carry on sixty hires a year, that leaves the work manual.

Greenhouse

AI features are marketed as benefits — a notetaker, voice AI, talent matching, AI-powered report filters — with the contract stating the customer remains solely responsible for hiring decisions and the vendor publishing AI principles plus a bias audit statement. But the captured pages do not document what each feature evaluates, offer a per-feature switch, or name a model provider, and the privacy policy discloses training AI models on personal information — that is a match score I could not explain to a candidate or a works council.

Workable

The agent can be disabled across the account at any time, configured per job down to screening thresholds and scoring weights, every score comes with its reasoning shown, protected characteristics are stated to be excluded by design, and low-confidence inferences pass to a human — the controls I need before showing a works council anything. The providers behind the acknowledged third-party large language models are not named on the captured pages, and I found no public information on informing candidates that AI is used or on EU processing for the AI features. The captured pages also give different pictures of automation: the privacy policy states no automated decision-making, while product pages say strong fits auto-advance with instant override.

Greenhouse

An API with usage limits, free data export at any time during the term, SSO in the entry plan and a developer sandbox with sync in the top tier are real substance my two-person team could build on. But the 'hundreds of pre-built integrations' are unnamed, and I found no public information on a handover to any HR or payroll system common in Germany, on e-signature, or on webhooks.

Workable

A fully documented REST API with webhooks, single sign-on, e-signatures, and an automatic export the moment a candidate reaches the hired stage — carrying the resume PDF, start date and personal email — plus the vendor's own HR and payroll-preparation modules means a hire does not get retyped. I found no public information on a maintained handover to HR or payroll systems common in DACH such as Personio or DATEV-connected systems, on SCIM, or on published API rate limits; the payroll integrations are named without naming the systems.

Greenhouse

The contracting entity is Greenhouse Software, Inc. of New York under New York law, the privacy policy says primary offices are in the United States and personal information may be processed in any of these countries, and hosting is unstated. The EU-U.S. Data Privacy Framework certification is a transfer safeguard, but I found no published subprocessor list and no statement of where the AI or any CV parsing runs — and every applicant record belongs to a third party who never chose this vendor.

Workable

The contracting entity is a UK company under English law and the exclusive jurisdiction of English courts, and the services are hosted and operated in the United States through Workable, Inc., with the US as the stated default and EEA storage only mentioned as a possibility — I found no public information on a customer-selectable EU region. A sub-processor list is published and transfers rest on SCCs and the Data Privacy Frameworks, which softens but does not change that every CV belongs to a candidate who never chose this vendor and is heading into a US-hosted stack.

Greenhouse

The pricing page names three tiers and shows which features sit in each, and the contract states fees are based on employee headcount including affiliates, due in advance, exclusive of taxes, on an automatic one-year renewal with a fee increase possible. But I found no public price figures for any tier, and API overages and professional services are charged at then-current rates, so the annual invoice is not computable from the captured pages.

Workable

All three bundle tiers are priced publicly, add-ons are itemised — texting at $89/mo, video interviews at $109/mo, assessments at $59/mo — and AI credits are priced per bundle down to one credit per candidate screened, which is more visible than most vendors manage. I found no public information on what the tier price is based on — employees, jobs or seats — on job-board posting costs or the booking model, on the minimum term, or on VAT treatment, so I cannot compute the real annual invoice for my volume. The 5% renewal cap with 45 days' notice is a useful published guardrail.