Serving Citizens Who Search in Bahasa Gaul

By Tessar Napitupulu, Founder of PT Arfadia Digital Indonesia, established 2008.

A government page written entirely in formal, official Bahasa Indonesia is structurally invisible to the citizens who search in colloquial, code-switched phrasing like "gimana cara buat NPWP online 2026", even when that page contains the exact information they need. The fix is not choosing one register over the other. It's running both, deliberately, on the same page.

Two Registers, One Ministry

Indonesian government content typically defaults to formal terminology: "prosedur permohonan paspor secara daring," "persyaratan pengajuan izin usaha mikro kecil." That register is correct for a legal document and nearly invisible in search behaviour, because citizens rarely type the way regulations are written. They type "cara buat paspor online 2026," "syarat izin UMKM online," often blending formal and informal Indonesian in the same sentence, "gimana cara" (colloquial) followed by a formal noun phrase.

Linguistic research into Indonesian search behaviour has documented this pattern extensively outside government content too: a single common term like "handphone" carries at minimum a dozen significant search variations across formal and informal registers. The choice between "Anda" and "kamu" or "lu" affects not just tone but signals of trust and intent, which matters specifically for YMYL content like government services, where formal "Anda" language tends to align better with what searchers expect from an authoritative source, even as their own query uses informal phrasing.

The practical objection to running both registers is usually workflow, not disagreement about whether it works. Formal content already has to survive an approval chain built around policy accuracy. Adding a colloquial FAQ layer means either training the same reviewers to sign off on informal phrasing, which can feel like a step backward to a communications team used to formal register as a proxy for correctness, or adding a second, lighter review step specifically for the FAQ content, since it carries the same factual weight without the same tone requirements. Neither option is complicated, but both require someone to actually decide which model applies, rather than defaulting to formal-only because that's what the existing workflow already handles.

Getting this right also means resisting the instinct to write the colloquial layer as a simplified paraphrase of the formal one. Citizens searching "gimana cara buat NPWP online 2026" are not asking for a shorter version of "prosedur permohonan NPWP secara daring." They're asking the same question in the words they'd actually use out loud, which is a different writing exercise than compression. The FAQ heading should, in many cases, use that exact colloquial phrasing verbatim, not a cleaned-up approximation of it.

The Government Calendar Nobody Plans Content Around

Citizen-service search volume in Indonesia is not evenly distributed across the year. It spikes predictably around a small set of recurring government events, and because content approval chains in government routinely run two to six weeks for anything substantive, planning has to lead these dates by a comfortable margin, not scramble to catch up once the spike has already started.

Event	Search Window	Representative Query
SPT Tahunan (annual tax return)	February to March, ahead of the April 30 deadline	"batas waktu lapor pajak 2026"
PPDB school enrollment	May to June	"jalur zonasi PPDB 2026"
Passport and travel document renewal	May to August, tied to school holidays	"cara buat paspor online 2026"
THR compliance (Lebaran)	6 to 8 weeks before Lebaran	"perhitungan THR 2026"
BPJS Kesehatan/Ketenagakerjaan enrollment	Rolling, spikes with announcement cycles	"cara daftar BPJS mandiri"
Pilkada/Pemilu cycles	6 months before election date	"jadwal pemilu daerah"
CPNS/PPPK recruitment	August to November	"formasi CPNS kementerian", "latihan soal CAT", "batas usia pendaftaran PPPK"
Bantuan sosial (PKH/BLT)	Rolling, spikes with each disbursement announcement	"kapan bansos pkh cair", "daftar nama penerima bansos"

Two of these categories deserve more attention than they usually get, because the search behaviour is urgent and the stakes are high for the citizen on the other end. CPNS and PPPK recruitment queries spike hard between August and November, driven by candidates checking formation quotas, practice tests for the CAT (Computer Assisted Test) exam, and age eligibility limits, often under real time pressure with a registration deadline days away. Bansos queries, checking when PKH or BLT disbursements land and whether a specific name appears on a recipient list, follow no fixed calendar at all: they spike the moment an announcement drops, which means the content answering them needs to already exist and be structured before that moment, not get written reactively once the search volume is already climbing. A ministry that only plans content for the predictable calendar items misses the rolling, announcement-triggered category entirely, and that category disproportionately serves citizens who may have fewer alternative ways to get the information quickly and accurately.

BUMN content follows a related but distinct pattern worth naming separately. Queries like "lowongan kerja bumn terbaru" and "program csr bumn untuk umkm" capture career and partnership-opportunity search intent rather than citizen-service intent, and they matter for a different reason: BUMN success metrics for content tend to track differently from ministry metrics, closer to brand reputation and stakeholder engagement than call-centre deflection, which changes what "good" content looks like even when the underlying language-register problem, formal institutional phrasing losing to how people actually search, is identical.

None of these dates are a surprise. Every one is knowable a year in advance, which is exactly what makes reactive content planning, starting the update once the search spike is already visible in analytics, an avoidable failure rather than an unpredictable one.

The reason it happens anyway is rarely a lack of awareness. Communications teams generally know when PPDB season or SPT Tahunan deadlines fall. What breaks the plan is the approval chain itself: a content update that needs sign-off from a technical bureau, then Kepala Humas, then in some cases the Secretary General, competing for the same limited review capacity every other piece of substantive content is also competing for. Seasonal content only gets prioritised ahead of the spike if someone has explicitly flagged it as time-sensitive months in advance, which requires the calendar existing as a document someone actually checks, not institutional memory that assumes everyone already knows.

Where Accessibility and Language Overlap

The dual-register approach solves a search problem, but it sits next to a second, related obligation: making the same content genuinely accessible. A study across all 34 Indonesian provincial government websites found 2,088 accessibility violations against WCAG 2.1, concentrated in the Perceivable principle, contrast, text alternatives, captions. Plain-language, well-structured content built for the colloquial-search layer often satisfies both goals at once: shorter sentences, clearer headings and explicit lists are easier for a citizen with a cognitive or visual disability to parse, and easier for an AI crawler to extract, for the same underlying reason.

Indonesia lacks a single binding national WCAG mandate as of 2025, but Presidential Instruction No. 3/2003, UU No. 8 of 2016 on the rights of persons with disabilities, and UU No. 1 of 2024 together create real legal exposure for non-compliant agencies, independent of whatever SEO or GEO benefit accessibility remediation happens to produce alongside it.

Progress on this front is measurable, not just aspirational. One practical formulation, an Accessibility Optimization Index, expresses improvement as a simple percentage: one minus the ratio of remaining violations to the baseline violation count found in an initial automated audit, multiplied by one hundred. A ministry that starts with 200 flagged violations from a tool like aXe DevTools and remediates down to 40 has moved its index to 80 percent, a number that translates cleanly into a progress report a non-technical director can actually follow, which matters when accessibility work competes for budget against initiatives with more obviously legible metrics.

The overlap is worth naming explicitly because it changes how a limited budget should be allocated. A communications unit weighing whether to spend its remaining quarter's budget on a language-and-search initiative or an accessibility remediation project is often choosing between two line items that, done correctly, are substantially the same underlying work: clearer headings, shorter sentences, explicit lists, properly labelled form fields. Treating them as separate initiatives with separate budgets usually means paying twice for overlapping deliverables, once under an accessibility compliance line and once under a digital communications line, when a single well-scoped project could satisfy both mandates at once.

The UU KIP Disclosure Calendar as Content Strategy

Law No. 14 of 2008 on Keterbukaan Informasi Publik creates three content-relevant obligations that map directly onto a working content calendar rather than sitting as an abstract legal requirement. Berkala, periodic information, must be published on a regular schedule. Serta-merta, immediate information, must go out rapidly when public safety or rights are affected, functionally a breaking-news content category. Setiap saat, at-all-times information, must be available on request, mapping to full service pages and FAQs kept permanently current.

Treating these three categories as a content calendar, rather than a legal filing obligation handled separately from the website, closes a compliance gap and an SEO gap with the same piece of work. A ministry that is behind on berkala disclosures is, by definition, also behind on the searchable, structured content an AI engine or a citizen would need to find that same information.

Ownership matters more than the calendar itself. A UU KIP disclosure calendar that lives only inside a legal or compliance unit's tracking spreadsheet, disconnected from whoever manages the actual website, tends to produce disclosures that are technically published but poorly structured, satisfying the letter of the law while remaining nearly as hard for a citizen or an AI engine to find as if they'd never been posted. The version that actually closes both gaps has one shared calendar, visible to both the compliance owner and the content owner, with each disclosure treated as a publishing task, not just a legal checkbox.

Answer Capsules, Not Just Answers

Writing the colloquial FAQ layer well is a specific skill, not a simplified version of the formal skill a communications officer already has. One useful discipline borrowed from structured GEO methodology is the "answer capsule": a self-contained response under 40 words that fully resolves the question on its own, without requiring the reader to scroll down for context. "Kapan bansos PKH cair" gets answered in the first sentence with a date or a clear "check here" mechanism, not after two paragraphs of programme background. The programme background can still exist further down the page for the reader who wants it. It just can't be the thing standing between the citizen and the one fact they came for.

Getting communications staff to write this way consistently is a training problem as much as a writing-guidelines problem. Aparatur Sipil Negara working in pranata humas roles are typically trained in formal public communication, press releases, and stakeholder messaging, not in writing for extraction by a machine or for a citizen who wants the shortest possible path to an answer. Structured technical guidance sessions (bimbingan teknis, commonly shortened to Bimtek), covering both the mechanics of dual-register writing and how to use AI drafting assistants without losing accuracy or the service-oriented tone government communication requires, close that gap faster than a written style guide alone, because the skill is closer to a habit than a rule that can be memorised from a document.

For a fuller treatment of programmatic, Indonesia-specific content strategy, including how entity SEO and structured data scale across large content sets like this, Tessar Napitupulu's Found Before They Search covers the model this article draws from, free to read. Agencies mapping their own seasonal and disclosure calendar against actual search behaviour can start with Arfadia's Government SEO service, and the companion piece on what changes once AI engines are answering citizen questions directly covers what happens after this foundation is in place.

---

Frequently Asked Questions

Should we rewrite our formal content entirely into colloquial language?

No. The formal, canonical version should stay intact for legal and compliance purposes. The colloquial layer sits alongside it, typically in the FAQ section, rather than replacing the official register anywhere.

How far in advance should we plan content for these seasonal spikes?

At minimum eight to ten weeks ahead of the spike, given typical government content approval timelines of two to six weeks for substantive pieces, plus buffer for review and publishing logistics.

Does this seasonal approach apply to BUMN content as well as ministries?

Less directly. BUMN content tends to follow disclosure schedules and corporate communications cycles rather than the citizen-service calendar described here, though any BUMN offering citizen-facing services, like BPJS Ketenagakerjaan, faces a similar seasonal pattern.

Is there a risk that colloquial language undermines the institution's authority?

Not if it's confined to the FAQ and heading layer while the substantive answer remains accurate and properly sourced. The authority signal comes from correctness and structure, not from formality of tone alone.

How does UU KIP compliance actually get measured?

Through periodic evaluations, typically conducted at the provincial or institutional level, checking whether required berkala disclosures were published on schedule. The 77.5% figure cited above comes from one such evaluation in Banten Province, not a national aggregate.

---