Boon Referral Program Implementation: Controls and Tests
On this page 13 sections
Boon is a recruiting referral platform at goboon.co that supports employee and external referrals, candidate matching, communications, rewards, and ATS-connected workflows. This article concerns that product, not unrelated companies using the name Boon. A successful implementation depends less on launching a branded portal than on defining who may refer whom, how attribution is decided, when rewards become payable, which data moves to the ATS, and how bias and privacy are monitored.
Start with one business unit, a small set of approved jobs, and a written referral policy. Keep the ATS as the system of record for applications and hiring decisions unless the architecture explicitly says otherwise. Do not pay rewards or trigger consequential candidate actions until identity, status, and attribution reconcile across systems.
Implementation verdict
Boon is worth piloting when referral participation is low because discovery, communication, or administration is fragmented. It cannot repair an unclear reward policy, inconsistent job eligibility, poor candidate follow-up, or a workforce network that reproduces narrow sourcing patterns.
The launch team should include talent acquisition, HR operations, HRIS or integration, payroll or accounts payable, privacy, security, legal, communications, and a business sponsor. Assign one owner for disputes.
Product boundary
Boon’s main product site describes employee referrals, automated matching and outreach, tracking, analytics, and rewards. Its external referral page extends the stated model beyond employees to outside participants. These are vendor descriptions, not evidence of outcomes for a particular employer.
Define which population and functions are in scope:
| Area | Policy decision | Product test |
|---|---|---|
| Referrer eligibility | Employees, alumni, contractors, external members, or selected groups | Unauthorized person cannot submit or see restricted jobs |
| Candidate eligibility | Prior applicants, former workers, relatives, vendors, or duplicates | Ineligible and duplicate cases follow written rules |
| Job eligibility | All roles or a controlled subset | Closed, confidential, and excluded roles are suppressed |
| Attribution | First referral, accepted referral, or ATS application source | Conflicts resolve consistently with an audit trail |
| Reward | Amount, milestone, tax treatment, approval, and forfeiture | Status and payment event reconcile to payroll or finance |
| Communication | Notice, consent, frequency, and opt-out | Suppression applies across Boon, ATS, email, and text |
Do not let defaults silently become company policy.
Map the systems of record
Document authoritative sources for workers, referrers, jobs, candidates, applications, hire status, employment start, reward approval, and payment. Boon may orchestrate referral engagement, but another system may own each underlying fact.
Boon’s integrations page names ATS and HR technology connections. A logo or connector name does not show direction, field coverage, frequency, error handling, or plan entitlement. Create an interface record for every object and event.
Use stable identifiers instead of names or emails alone. Rehires, aliases, personal email changes, multiple applications, and agency submissions make identity matching unreliable without clear rules.
Greenhouse example shows the setup boundary
Greenhouse publishes an independent Boon integration guide. It documents a specific connection from Greenhouse’s perspective and is stronger evidence for that integration than Boon’s general catalog. It does not prove behavior for another ATS or every Greenhouse configuration.
For any ATS, test job synchronization, close and reopen behavior, candidate creation, source, referrer identifier, application status, duplicate detection, hire event, start-date change, and deletion. Pause the integration, create events on both sides, restore it, and confirm ordered recovery without duplicates.
Name the support owner for a failed record. An interface that runs automatically but cannot be reconciled increases administrative risk.
Attribution rules before rewards
Referral disputes often arise when the candidate already exists, applies directly, is submitted by an agency, receives multiple referrals, or changes roles. Write precedence rules before launch. Include a time window, evidence required, manual review, appeal, and final authority.
The platform should retain the original referral, later applications, rule applied, decision, reviewer, and timestamp. Avoid overwriting source merely to match the eventual payout. Recruiting analytics and reward eligibility can use different fields if both are clearly defined.
Run at least these cases:
- Two employees refer the same person for one job.
- A candidate in the ATS is referred for a different job.
- An agency candidate later receives an employee referral.
- A referral occurs after a direct application.
- A hire changes start date or leaves before the reward milestone.
- A referrer changes employment status before payment.
Reward operations
Define whether a reward is cash, points, donation, recognition, or another benefit. Record currency, approval, tax treatment, eligibility date, waiting period, forfeiture, split, cancellation, and payment system.
Do not make Boon the final payroll authority unless that is a deliberate and tested design. Send an approved reward event with the required identifier, then reconcile accepted, rejected, paid, and reversed transactions. Restrict users who can change hire or reward status.
External referrers may require different identity, tax, payment, terms, and fraud controls. Do not extend an employee policy by changing only the login page.
Matching and AI controls
Boon’s AI page describes matching and automation. Treat every result as a vendor-generated recommendation. Ask what candidate and job fields are used, whether inferred data is included, how the model is evaluated, which provider processes data, how long inputs and outputs are retained, and whether the feature can be disabled.
Referral networks can concentrate opportunity within existing social and occupational groups. A matching score may add another selection layer. Compare referred and non-referred applicants by source, stage, job family, and relevant demographic measure where lawful. Investigate disparities rather than assuming referrals are representative.
The EEOC’s AI and disability resources explain that employers retain responsibility for avoiding discriminatory barriers when using software. The guidance does not evaluate Boon. Use it to test accommodations, alternative submission routes, and human review.
Candidate and referrer privacy
A referral may expose a person’s name, contact details, relationship, employer, or interest before the person has interacted with the company. Decide what a referrer may submit and when the candidate receives notice. Minimize free-text fields and prohibit sensitive details that are unnecessary for recruitment.
Test access, correction, withdrawal, suppression, retention, and deletion across Boon and the ATS. Decide what remains when a candidate declines interest or asks not to be contacted. Restrict referrers to an appropriate status view rather than revealing interview or decision details.
For external networks, publish terms and a privacy notice suited to that population. Confirm international transfers, subprocessors, messaging providers, and payment processors.
Security and contract evidence
Boon’s security page provides the vendor’s current security statements. Request the evidence behind relevant claims: assurance reports or certificates, scope, audit period, exceptions, penetration-test summary, recovery objectives, incident notice, subprocessor list, and data deletion.
Test SSO and provisioning if included, administrator recovery, referrer role, recruiter role, reward approver, export rights, API credentials, audit logs, and terminated-user access. A former employee should not retain referral or reward visibility through an unmanaged personal account unless the external program explicitly permits it.
Boon also publishes trial terms. Review trial data, conversion, cancellation, export, deletion, confidentiality, and publicity before loading candidate or employee information. A trial agreement may differ from the eventual subscription, so attach the controlling production terms to the final order.
Four-week pilot
| Week | Work | Pass condition |
|---|---|---|
| 1 | Policy, roles, eligible jobs, notices, identifiers, and integrations | Owners approve the written design |
| 2 | Employee and candidate invitation, referral, duplicate, and opt-out cases | Identity, consent, and communications remain traceable |
| 3 | ATS outage, attribution conflicts, hire changes, reward approval, and reversal | Exceptions recover and financial events reconcile |
| 4 | Security, privacy, bias monitoring, support, export, and commercial review | Evidence package is accepted and unresolved items are contractual |
Use synthetic candidates unless real people have been informed and authorized for the pilot. Do not measure success only by referral count. Track eligible participation, accepted referrals, candidate response, process time, disputes, duplicate rate, adverse-impact indicators, reward reconciliation, and opt-outs.
Launch acceptance matrix
| Test | Pass condition |
|---|---|
| Job synchronization | Only eligible open roles appear and closures propagate on time |
| Identity | Referrer and candidate records use stable IDs and duplicates enter review |
| Attribution | Every conflict resolves from written precedence with an audit event |
| Communication | Notice, frequency, unsubscribe, withdrawal, and suppression work across systems |
| Hiring status | ATS remains authoritative and delayed events do not trigger false rewards |
| Reward | Eligibility, approval, tax handoff, payment, and reversal reconcile |
| Access | Employees, external referrers, recruiters, approvers, and administrators see only allowed data |
| Exit | Referrals, relationships, statuses, approvals, and audit history export in usable form |
Require the production configuration and proposed contract to produce the result.
Unknowns to close
Public pages do not establish a universal implementation time, final price, connector behavior, model accuracy, data location, support response, reward-tax treatment, or measurable hiring outcome. Vendor outcome metrics should not be applied to a new employer without its own baseline and measurement design.
Request a feature schedule, itemized quote, implementation statement, integration specification, data-flow diagram, AI inventory, assurance package, subprocessor list, service levels, trial and production terms, sample export, and deletion confirmation. Identify any custom work and its owner.
Go-live decision
Launch Boon when referral eligibility, identity, candidate notice, attribution, ATS synchronization, rewards, access, monitoring, and export all pass. Delay when the program lacks a dispute rule, hire events cannot reconcile, external referrers are governed by employee assumptions, or automated matching has no documented human review. A referral platform amplifies the program it is given; policy and controls determine whether that amplification is useful.