OTAs Own Generic Search, Not the Booking

By Tessar Napitupulu, Founder & CEO, PT Arfadia Digital Indonesia, GEO pioneer in Indonesia since 2023. More from Tessar.

Traveloka and Tiket.com will win almost any search for "paket wisata bali" or "hotel di lombok." That fight is not worth having. The search terms that actually convert, the ones with a destination, a duration, and a specific experience attached, are the ones an independent tour operator can still win outright, and mostly nobody is bothering to build for them yet.

Indonesia's domestic travel market gives this fight real stakes. BPS recorded roughly 1.2 billion domestic tourist trips in 2025, up 17.55% from 2024 and the highest figure in seven years, comfortably above the 722.15 million pre-pandemic baseline. That volume has to search for something before it books, and where it searches is splitting cleanly into two lanes: a generic lane the OTAs have owned for a decade, and a specific lane almost nobody has claimed.

Why Can't Tour Operators Just Outrank Traveloka?

Traveloka holds roughly 45% of Indonesia's OTA market as of 2025, and Traveloka plus Tiket.com together were estimated at around 72% of the country's online travel booking value back in 2022, split roughly 51% and 21%. That combined figure gets repeated constantly in industry decks, and it is worth flagging plainly: it is a dated 2022 estimate of booking value, not a current measurement of search visibility, and the two things are not the same claim. Even with that caveat, the direction is not in doubt. A decade of domain authority, thousands of destination pages, and advertising budgets no independent operator can match make the first page of results for a generic destination term effectively spoken for.

The practical answer is not to compete there. It is to walk the funnel down to where OTAs stop being able to compete at all.

The Four Stages a Traveler Actually Searches Through

Travel search intent in Indonesia moves through four distinct stages, and each one calls for a different kind of page, not a different keyword on the same page.

• Destination inspiration (six to twelve months out): "tempat wisata terbaik di indonesia," "destinasi liburan romantis." High volume, low commercial intent, heavily exposed to AI Overviews summarizing the answer directly.
• Itinerary research (three to six months out): "paket wisata bali 5 hari 4 malam," "itinerary raja ampat 7 hari." Active planning intent, and the single highest-priority commercial cluster for a tour operator.
• Comparison (one to three months out): "open trip vs private trip bromo," weighing the DIY-versus-organized-tour decision directly.
• Booking (days to weeks out): "pesan paket wisata bali murah terpercaya." Fully transactional, and the only stage that should carry a hard price and a checkout path.

Conflating these four into a single page is the single most common structural mistake in travel SEO. An inspiration page trying to also sell a package usually inspires poorly and sells worse, because it is answering none of the four questions cleanly.

A funnel documented in general travel SEO practice shows how a single traveler's search actually evolves across these stages, even outside the Indonesian market specifically: a search for "maui snorkeling" (pure research) becomes "best snorkeling tour maui price" (comparison) and finally "book snorkeling tour maui" (transactional), often over several separate sessions. Attribution models that only credit the final click badly undercount the research-stage content that started the journey. An operator without an inspiration or comparison page in that funnel is invisible for the two stages where the eventual buyer was actually forming a decision, and only shows up, if at all, at the very last step where competition is highest.

How Is the Shift to AI Search Changing the Calculus?

Google's AI Overviews recorded a 381% increase in appearances for travel queries during the March 2025 core update alone, one of the sharpest category-specific spikes measured anywhere in search. A Seer Interactive study published in late 2025 found organic click-through rates on queries with an AI Overview present dropped 61%, from 1.76% to 0.61%, while paid CTR fell 68% over the same period. That looks like uniformly bad news for anyone relying on organic search.

The same dataset contains the more important half of the story: brands cited within an AI Overview earned 35% more organic clicks and 91% more paid clicks than brands not cited, on the exact same query set. The competitive question for 2026 has shifted from "do we rank in the top five" to "are we the source the AI is drawing its answer from." A long-tail itinerary page built with real prices, named locations, and validated schema is not just optimizing for a shrinking organic-click pool, it is auditioning for the citation slot that determines whether that click pool shrinks around you or past you.

Which Schema Types Actually Move the Needle

Five schema types cover the practical deployment priority for a travel operator's site, in the order they should go live: Organization, to establish the business as a distinct entity in Google's Knowledge Graph; TouristTrip, on every per-tour page, with the itinerary property populated as an ItemList of sequenced Place stops; AggregateRating, sourced only from verified reviews, associated directly with the CTR uplift vendors commonly cite in the 20 to 35% range for rich-snippet display; FAQPage, on the exact questions and answers that are visibly on the page, not hidden text; and BreadcrumbList, for navigation hierarchy. One caveat worth stating plainly: schema markup does not guarantee a rich result or a citation. Google's own documentation says structured data helps a page's eligibility and machine understanding, not its guaranteed display. What deploying it correctly does is remove a page from consideration entirely if it is missing, which is a much larger risk than most operators realize.

What Does the Winning Keyword Pattern Actually Look Like?

The "paket wisata [destinasi] [durasi]" pattern is the structurally strongest format available to an Indonesian tour operator, and it is strongest for a specific reason: OTAs cannot build destination-depth content at this level of specificity without abandoning the aggregation model that makes them OTAs in the first place. A page for "paket wisata lombok 5d4n private snorkeling gili" has far lower search volume than "paket wisata lombok," and dramatically less competition, because almost none of that competition is an OTA.

A documented case from a Malang-based tour operator makes the mechanism concrete: separating a general "paket wisata bromo" page from a price-sensitive "paket wisata bromo murah" variant, then building a topical cluster around each covering itinerary routes, private-versus-open-trip comparison, safety information, and preparation checklists, produced measurable ranking and click-through improvements. The operational rule that generalizes from this: each destination an operator actually serves should carry a cluster of four to six supporting articles linking to one central pillar page, not a single page trying to cover the whole topic alone.

What Can an Operator Do That an OTA Structurally Cannot?

Five content dimensions separate a defensible operator page from an aggregator listing, and none of them require competing on the aggregator's own terms.

Dimension	Why an OTA can't replicate it
First-hand experience content	Field photography and guided-tour notes from an actual departure, not stock imagery.
Price transparency with breakdown	Itemised transport, accommodation, guiding fee, and park entry costs answer "is this fair," which a headline price alone doesn't.
Private vs. open trip comparison	An honest classification of tour modalities, a content asset an aggregator has no incentive to build.
Local safety and logistics depth	Visa rules, health certifications, and emergency contacts specific to a route, not a generic disclaimer.
Structured review architecture	AggregateRating schema on a specific tour page, tied to that exact itinerary rather than a general brand score.

A useful proof point for the ceiling on this strategy: a July 2026 study of 105 Bahasa Indonesia travel queries in Google AI Overviews found citations spread across 298 unique domains, not concentrated among a handful of high-authority sites. Smaller operators with well-structured, experience-specific content earned citations ahead of platforms with far higher domain authority. The field genuinely is open, not just in theory.

Does This Mean Traditional SEO Metrics Still Matter?

Some do, and some are actively misleading for a tour operator. Standalone keyword rankings, exact-match keyword presence, and raw page-view volume without engagement context should all be deprioritized. Tour bookings are considered purchases with decision timelines running days to weeks, so a destination-inspiration article generating ten thousand monthly visits and two package inquiries is not outperforming a targeted itinerary page generating three hundred visits and fifteen inquiries. It is underperforming it, badly, and a traffic-only dashboard will never show that.

The metrics that hold up: package inquiry submission rate as the primary conversion KPI, inquiry-to-booking close rate as a separate measurement of sales-process effectiveness (industry benchmarks put this at roughly 15% for high-intent traffic), and AI Overview citation presence as an emerging visibility signal that Google Search Console does not yet isolate on its own.

Where Should an Operator Start This Quarter?

Pick two or three destinations where margin, capacity, and internal expertise are strongest, not every destination the operator technically services. Build one full hub-and-spoke cluster for each: a pillar page, four to six supporting articles, TouristTrip and FAQPage schema validated against Google's Rich Results Test, and a dedicated operator-credibility page linked from every package. Measure inquiries and inquiry-to-booking rate from week one, not traffic. Expand to the next destination only once that cluster is actually converting, not once it is merely ranking.

The generic search terms were never winnable, and chasing them wastes a content budget that could be building the long tail instead. The itinerary and package searches that book are still open, and the operators who structure content to win them are compounding an advantage that gets harder to catch every quarter they wait.

This is the SEO half of a two-part strategy. The other half, structuring the same content so AI systems like ChatGPT, Perplexity, and Google AI Overviews cite it directly, is covered in our GEO for travel service, and in the piece on how Instagram is already out-citing OTAs in Indonesian AI search.

---

Frequently Asked Questions

Should we build a separate page for every destination we serve, even small ones?

Only where there is real margin and capacity to fulfil bookings. A pillar-and-spoke cluster takes real content investment. Building one shallow page per destination for coverage's sake produces thin content that neither ranks nor converts.

Is the "paket wisata [destinasi] [durasi]" pattern still worth targeting if search volume looks low?

Yes, and the low absolute volume is part of why it works. Lower volume with dramatically lower competition and higher booking intent outperforms high volume with near-zero realistic ranking chance against an OTA.

How many supporting articles does one destination cluster actually need?

Four to six is the range that shows up consistently in practitioner case studies: itinerary routes, private-versus-open-trip comparison, safety and preparation, and often a halal or visa angle depending on the destination.

Do we need real search-volume data before committing to specific target keywords?

Yes. Actual Indonesian search volumes for these long-tail queries are not reliably published; they need to come from a keyword tool like Ahrefs or Semrush before a client-facing target is finalized, not estimated from general practitioner commentary.

What happens to inspiration content in this model? Should we stop writing it?

No, but stop measuring it by its own traffic. Its job is to feed the funnel down toward itinerary and package pages. Measure it by assisted conversions and internal click-through into the cluster, not standalone sessions.

Does the shift toward AI Overviews mean traditional SEO is no longer worth the investment?

The opposite. Content built with real prices, named locations, and validated schema is exactly what an AI Overview needs to extract in order to cite a page. Traditional SEO discipline and AI citation eligibility are increasingly the same underlying work, not competing priorities.

---