The Problem
Noven, 39, had been looking for formal work since 2022 while earning money as an app-based motorcycle driver. At a Jakarta job fair, he told detikFinance that company age requirements were getting in his way. Raka, a Deaf applicant, described a different barrier: an employer appeared to offer an administration vacancy, then said it could not accept an applicant with a disability. Another applicant, Suci, wanted employers to stop leaving her without an answer. These accounts appeared in reporting published on 24 October 2026. [1]
These are individual experiences, not a measure of how often Indonesian employers exclude applicants. They nevertheless pose a practical question: how can someone get a fair hearing when their application is screened out before their ability is considered—or disappears without a useful reply?
There are resources to build on. Jakarta’s regional job-training centres, known as Pusat Pelatihan Kerja Daerah (PPKD), already teach vocational skills; the Central Jakarta centre says it works with employers on placing graduates. The city also operates KarirJakarta, a public job portal with links to disability vacancies and training. [2] [4] The opportunity is to improve what happens between those services and an employer’s decision.
What Could Work?
Five possibilities rearrange different parts of that journey. All are untested proposals; the suggested buyers and their willingness to pay remain unverified.
- Skills-First Applications: let participating employers examine a short work sample before a conventional résumé. An employer might pay a service operator per vacancy to prepare and route those applications, if this produces useful interviews without excessive screening time.
- Fair Skills Booths: give fair visitors equipment, an accessible task and a trainer who can connect the result to participating recruiters. A city labour office might pay an operator for an assisted assessment station if it improves interview access beyond ordinary fair services.
- Feedback-Linked Posting: make access to a participating portal’s special posting lane depend on giving applicants timely, substantive replies. The portal might buy an integration service and charge employers per post if accountable communication attracts applicants without driving employers away.
- Coursework-to-Shortlist: turn an existing training assessment into a candidate-approved work sample that an instructor sends to employers with relevant vacancies. Employers might pay a coordinator for a curated feed if it saves duplicate assessment and sourcing effort.
- Paid Work Observation: add structured, supervised observation alongside a host employee, with compensation and a lawful arrangement settled before anyone starts. Employers might pay a coordinator per participant if observation answers questions a short test cannot.
How It Could Work
Skills-First Applications
For an administration vacancy, an employer and trainer could agree on a roughly 20-minute fictional order-checking task. Applicants receive the instructions, scoring criteria and compensation terms in advance. A coordinator verifies authorship, then sends coded outputs and scores to the recruiter before names, ages or employment histories. The employer records its shortlist decision before receiving the contact details needed for an interview.
This removes some early screening information and puts a relevant piece of work in its place. It does not make judgement bias-free. A separate coordinator must arrange accessibility support from the outset; candidates should not have to conceal a disability to request it. Reviewers see only information needed to judge the task.
Fair Skills Booths
A visitor without a suitable device could complete a narrower, roughly ten-minute task on a booth computer with a trainer nearby. The visitor receives specific feedback and can choose which participating recruiter receives a time-limited link to the result. A printed reference can serve visitors who do not want to use a scannable code.
The distinctive resource is assisted access at the fair: equipment, explanation and a recruiter are close together. The result would describe performance on one task, not certify overall employability. Recruiters would need to agree beforehand to consider it; a link alone cannot create an interview or verify an employer’s legitimacy.
Feedback-Linked Posting
A portal could invite employers into a lane where each vacancy has a published skills rubric. Within 14 days of an application, the employer must give either a decision with a criterion-based explanation or a genuine progress update with a next-update date. Closing the application requires a final reason. Missing the initial reply or a promised update date pauses new posts in that lane until the employer resolves the overdue replies, with a correction and appeal process for mistakes.
The inventive change is the dependency: continued posting access depends on communication. Existing status tools are a starting point—Indeed already describes employer-updated application statuses. [7] The proposed addition is an enforceable service commitment with checks on whether replies actually help applicants understand what happened. It cannot by itself make a rejection fair.
Coursework-to-Shortlist
Instead of asking a graduate to take another test, an instructor could retain an agreed piece of final coursework, its rubric and the date of assessment. With the graduate’s permission, a coordinator sends a small weekly batch to employers with matching open roles. The candidate confirms availability and approves each destination.
Here, assessment already done for teaching gains a second use in hiring. For an administration role, a redacted spreadsheet or screen-only recording may suffice. Face and voice video would undermine an initially masked review; identifying filenames, dates and other revealing details also need attention. The instructor can attest to how the work was produced, while the employer remains responsible for selection.
Paid Work Observation
A short task may reveal accuracy but miss how someone responds to instructions, correction or teamwork. A training centre and employer could explore three supervised observation days alongside an experienced employee, using a written checklist and a feedback meeting. Participants would be compensated, and tasks, access needs, supervision and protections agreed in advance.
This proposal has a prerequisite: labour counsel and the relevant labour office must establish the appropriate legal arrangement. Calling productive work an “assessment” does not settle its status. Indonesia’s domestic apprenticeship regulation has requirements for programmes, agreements, approval, allowances and social protection; it is not a blanket exemption for a three-day trial. [6] If no suitable arrangement can be secured, this pilot does not start.
What Could Get in the Way?
All five ideas depend on employers changing behaviour. A skills score is of little use if a recruiter later applies the same irrelevant exclusion. A 2025 manpower minister’s circular addresses recruitment discrimination and sets conditions on age requirements, including provisions applying to disabled workers. That policy is context, not evidence that these proposals will secure compliance. [3]
Skills-First Applications could add another barrier if the task is inaccessible, too burdensome or unrelated to the vacancy. A high score may not predict performance, and identifying clues can remain in supposedly masked work. Kalibrr already offers candidate assessment results, so the value would have to come from the agreed sequence and quality of review, rather than from simply having a test. [5]
Fair Skills Booths could create queues while recruiters remain too busy to act. Speed targets must not penalise people who need assistance or extra time. Indonesia’s disability law explicitly provides for adapted recruitment tests and time flexibility; accessible design belongs in the pilot from the beginning. [8]
Feedback-Linked Posting could reward empty templates or prompt employers to move elsewhere. A dated update can be honest without being a decision, but repeated updates must not become a new way to keep applicants waiting. Replies need correction and complaint routes, not assurances that a template removes legal risk.
Coursework-to-Shortlist depends on instructor time, comparable scoring and current employer demand. Coursework can become stale, and teaching assistance may make it a poor representation of independent work. It would also miss people outside participating courses.
Paid Work Observation requires host capacity and a lawful, protective arrangement. Three days may favour fast starters or expose candidates to avoidable travel costs and disruption. Observation must not become a rotating supply of cheap labour or a substitute for proper training.
Sharing candidate evidence also creates responsibilities. Indonesia’s Personal Data Protection Law requires a basis for processing and imposes security duties. Each operator would need a reviewed plan for purpose, recipients, access, retention, candidate rights and any sensitive information. A consent box alone does not settle those duties. [9]
What Would It Take to Find Out?
These are proposed feasibility tests, not results. Use real vacancies, volunteers aged 18 or older who meet the role’s relevant entry requirements, and accessible participation. Recruit people facing the barriers discussed here without treating age, disability or a long gap as evidence of inability. Compare each intervention with the same employer’s or service’s current process, and record people who decline or drop out. The thresholds below are provisional decisions to agree before recruitment, not estimates of likely benefits.
Skills-First Applications
Across three employers hiring for one administration role family, randomly offer 60 eligible applicants either the new route or the usual application route, 30 per group. Randomise within employers so each has both routes. Over 28 days, measure the share attending a substantive interview, counting everyone assigned, including task non-completers. An increase of at least ten percentage points would justify a larger test only if access complaints and screening time remain acceptable. No increase—or invitations that vanish after identity disclosure—would count against the mechanism.
Fair Skills Booths
At one two-day fair, randomly invite 40 of 80 eligible, consenting visitors to the booth; the other 40 use ordinary fair services. Compare attended interviews within 14 days for everyone assigned, not just those who finish a task. Look for at least a ten-percentage-point improvement, while measuring queues, assistance needs and cost per visitor assigned. Scans without interviews, or delays that stop visitors reaching recruiters, would argue against expanding it.
Feedback-Linked Posting
Recruit 20 willing employers on one portal and randomly assign ten to the posting rule and ten to current practice for eight weeks. Measure the share of applications receiving a substantive reply within 14 days, audit whether replies match the recruitment record, and survey whether applicants understand their status and next step. A 20-percentage-point improvement would be encouraging only if accuracy holds and no more than two of the ten intervention employers leave the lane. Generic replies or substantial exit would count against it.
Coursework-to-Shortlist
In one centre, randomly offer 40 consenting, job-ready graduates either the new evidence feed or the centre’s existing referrals, 20 per group. Send four weekly batches for the intervention group and follow both groups for 28 days from assignment. Compare attended-interview rates, instructor minutes per graduate and employer review time. Seek at least a ten-percentage-point interview improvement without exceeding a pre-agreed workload budget. Unopened batches or a need to retest everyone would undermine the proposed saving.
Paid Work Observation
Only after its legal prerequisites are met, two employers could recruit 20 eligible volunteers for comparable administration vacancies and randomly offer ten the observation route and ten the usual interview route. Over 30 days from assignment, compare written offers per assigned participant and record actual participation, compensation, host hours and complaints. At least two additional offers among the ten offered observation would justify further study only if the protections were delivered. No gain, burdensome hosting or any serious protection failure would stop expansion.
Small groups will produce uncertain estimates. Track subsequent job starts and, where participants agree, 90-day retention among those who start; report those denominators separately. None of these pilots can establish national impact or prove that an applicant who was not hired lacked ability.
Who Could Make It Work?
Skills-First Applications needs employers with genuine vacancies, a role specialist and a coordinator capable of checking work and controlling disclosure. Fair Skills Booths needs the fair organiser, trainers and recruiters willing to reserve interview capacity. Feedback-Linked Posting needs a portal product team, employer users and someone able to audit replies and handle appeals.
Coursework-to-Shortlist needs a training-centre placement team, instructors and employers willing to agree on useful evidence. Paid Work Observation needs host supervisors, a training coordinator, labour counsel and the relevant labour office before participants are recruited.
Across the tests, Deaf and other disabled job seekers should help design the process alongside an accessibility specialist and a data-protection adviser. None of the organisations mentioned here has agreed to participate. Which change would make employers more willing to consider an overlooked applicant—and which would merely give that applicant another task?
Research Notes
Sources checked on 24 October 2026. Reporting establishes what interviewees said; service documentation establishes described features; legal texts establish the cited provisions. None independently validates the proposed hiring outcomes. The interviews identify experiences worth investigating, without establishing their prevalence.
1. Curhatan Sulitnya Warga Dapat Kerja: Lamaran Tak Dilirik-Terganjal Usia
Original source. detikFinance; Ignacio Geordi Oswaldo; 24 October 2026. Original reporting, full article read. Supports the attributed accounts of Noven, Raka and Suci. These interviews do not measure national prevalence or employer demand.
2. PPKD Jakpus Siapkan Beragam Pelatihan Modern Bagi Warga Bersaing di Dunia Kerja dan Usaha
Original source. Central Jakarta city government; publication date not established from the article body. Full page read. The centre describes vocational training and employer cooperation for graduate placement. Permission to reuse individual assessment records, instructor capacity and participation by particular employers remain unverified.
3. Surat Edaran Menteri Ketenagakerjaan Nomor M/6/HK.04/V/2025
Original source. Ministry of Manpower; 28 May 2025. Both pages of the original scanned circular read visually; the PDF has no usable extracted text layer. Points 2–4 address discrimination, conditions on age requirements and application to disabled workers. The circular sets conditions on age requirements rather than announcing an unconditional removal of all age criteria. Enforcement was not assessed.
4. KarirJakarta public employment portal
Original source. Jakarta provincial employment service; live portal, no publication date stated. Full page read. Navigation links to disability vacancies, job fairs and regional training centres. These public channels are existing alternatives. Their placement effectiveness was not established.
5. How can I view the candidate’s assessment results?
Original source. Kalibrr Help Center; updated 9 June 2024. Full help article read through native web research; a direct download encountered an access challenge. Candidate-selected public assessment results already appear in recruiter profiles. Existing assessment tools are a relevant comparator for the proposed change in review sequence.
6. Peraturan Menteri Ketenagakerjaan Nomor 6 Tahun 2020: Penyelenggaraan Pemagangan di Dalam Negeri
Original source. Ministry of Manpower; established 7 April 2020, promulgated 9 April 2020. Original regulation downloaded; relevant provisions and surrounding requirements read, including Articles 3–5, 10–13, 16–17 and 22–24. The ministry’s regulation index marked it in force when checked. Agreements, programme and organiser approval, allowances, supervision, social protection and disability needs matter if the arrangement is an apprenticeship. Reading these rules does not establish the legal classification of a particular three-day activity; local advice and the required approvals remain prerequisites.
7. Lowongan Saya: Mengelola Lamaran yang Sudah Diajukan
Original source. Indeed Help Center; no publication date stated on the page read. Full page read. Describes application statuses, employer changes and limitations for off-site applications. The described status tools are an existing communication alternative; the proposed posting rule adds a service commitment.
8. Undang-Undang Nomor 8 Tahun 2016 tentang Penyandang Disabilitas
Original source. Republic of Indonesia, hosted by the Ministry of Manpower; 15 April 2016. Original law downloaded; recruitment and workplace provisions read in context. Article 47 provides for adapted recruitment tools and flexible test time; Article 50 requires reasonable accommodation and accessible facilities for disabled workers. Task design and support need to reflect these recruitment and workplace provisions.
9. Undang-Undang Nomor 27 Tahun 2022 tentang Pelindungan Data Pribadi
Original source. Republic of Indonesia, hosted by the Ministry of Finance; 17 October 2022. Original law downloaded; relevant processing, notice, consent, security and withdrawal provisions read. Articles 20–22 and 35–40 inform the proposed data review. Current implementing-rule status and the actual operator’s compliance are unverified. Before launch, check implementing requirements, any placement-service licensing, data responsibilities and public procurement rules for the proposed arrangement.
Skills-First Applications
Evidence and assumptions: assessment features are source checked [5]; the employer’s trust in this particular work sample, reduced premature exclusion, task validity and willingness to pay are unverified. Closest alternatives are existing portal assessments, employer work samples and ordinary résumé screening. The proposed difference is a recorded skills decision before identity disclosure, with separate access support.
Pilot and cost basis: use the 60-person, 28-day random offer described above, with one assignment per person and equivalent vacancy information in both arms. Report interview attendance as attended/assigned and the difference in percentage points; ten percentage points is three additional interviews out of 30. Report task completion separately, plus invitations and losses after disclosure. Employer fees would be per vacancy; operator costs include task design, scoring, candidate compensation, support and data handling. Cost per assigned applicant and cost per attended interview have different denominators and must not be compared as the same unit. Prices and net savings are unknown.
Fair Skills Booths
Evidence and assumptions: fairs and training infrastructure are source checked [1], [2], [4]; booth access, staffing, employer acceptance, budget and procurement approval are unverified. Alternatives include ordinary fair booths, public vacancy services and assessment tools employers bring themselves. Assisted completion and an immediate handoff are the proposed value to compare with existing fair services.
Pilot and cost basis: use 80 assigned visitors, 40 per arm, with the same eligible vacancies and follow-up at 14 days. Ten percentage points equals four additional interviews out of 40. Track task completion, queue departures, recruiter availability and accessibility separately. As an illustrative capacity check, one assessor working six productive hours daily over two days has 720 minutes; 15-minute slots allow at most 48 assessments before extra accommodations and disruption. Plan for 40 invitations and provide extra time or staffing as needed, rather than cutting access support. The proposed sale unit is a station for one fair; costs include staffing, equipment, venue arrangements, support and follow-up. Track cost/assigned visitor and cost/attended interview separately; no price is established.
Feedback-Linked Posting
Evidence and assumptions: existing statuses are source checked [7]; portal enforcement rights, integration access, employer retention, useful explanations and paid demand are unverified. Alternatives are status pages, recruiter messages and internal applicant-tracking templates. The proposed dependency applies only to a voluntarily joined lane, not a new legal requirement.
Pilot and cost basis: randomise employers, not individual applications; report results by employer as well as pooled application counts so a high-volume firm cannot dominate. Include applications received during the first six weeks of the eight-week pilot so each has 14 days of follow-up. A substantive progress reply must describe the actual stage and a next-update date; a rejection must identify a relevant published criterion. Audit a sample against records and separately report decisions, updates, overdue final closure, candidate understanding and complaints. Record application volumes, missed promised updates and employer exits alongside the reply-rate comparison. The proposed commercial unit is an employer post; integration, support and audit costs divided by paid posts determine cost per post. Price and legal review of the actual reply process remain open.
Coursework-to-Shortlist
Evidence and assumptions: training and placement partnerships are source checked [2]; record reuse permission, authorship verification, task comparability, vacancy matches and recurring employer attention are unverified. The strongest comparator is the centre’s own referral process, including certificates and instructor references. The proposed difference is reusing specific work already produced, not merely sending referrals weekly.
Pilot and cost basis: use 40 graduates, 20 per arm, and the same 28-day window from assignment, not a fresh window for each batch. Ten percentage points equals two additional interviews out of 20. Record all assigned graduates, even if no employer opens a sample. Document teaching assistance and assessment dates; audit scoring consistency. Agree the maximum added instructor minutes per graduate before enrolment and reject an uneconomic workload even if interview counts rise. Employers might pay per monthly feed; allocate curation, permission management and support costs across paying employer-months. Separately measure marginal instructor minutes/graduate. Unknown prices and demand prevent a profit estimate.
Paid Work Observation
Evidence and assumptions: apprenticeship requirements are source checked [6]; the legal route, host approval, compensation and protection arrangements for this specific pilot are unverified prerequisites. Alternatives include ordinary interviews, simulations, established apprenticeships and employment with lawful onboarding arrangements. The distinctive hypothesis is that structured observation with a host reveals relevant work behaviour missed by a brief task.
Pilot and cost basis: recruit only after the legal route and protections are confirmed. Use 20 assigned volunteers, ten per arm, and count written offers/assigned participant at day 30; two additional offers is a 20-percentage-point difference. Track starts separately from offers. A later retention rate uses job starters as its denominator and is subject to selection, not proof of reduced bad hires. Audit compensation, supervision, accessibility, host hours and complaints; any serious protection failure stops expansion. Employers would fund the coordinator and participant costs under the agreed arrangement. Before recruitment, cost compensation, protection, supervision and administration. Report costs per participant offered observation and per participant completing it separately.
All numerical thresholds and pilot sizes above are proposed design choices. Small samples and willing employers limit inference; report raw counts and uncertainty, then redesign or expand only if the specific mechanism shows promise.
Leave a Reply