Changi Airport boosts search with OpenSearch overhaul

Changi Airport Group has deployed OpenSearch across its website and mobile app to unify search for airport services, retail and travel information. The aim is to make it easier for passengers to find information and products across Changi's digital channels.

The new system brings together results for flight information, shopping and dining options, airport facilities, events and listings from iShopChangi, its eCommerce platform. It replaces a legacy setup that relied heavily on keyword matching and exact-term tagging.

That older approach often struggled when travellers used terms that did not match indexed wording. Searches for everyday terms such as ATM could fail if the system recognised only "Cash Machines", creating dead ends for users looking for services inside the airport.

As Changi's digital estate expanded across terminals and Jewel, the group wanted search to do more than function as a basic lookup tool. It wanted a broader discovery layer. OpenSearch gave the internal team direct control over search architecture, data ingestion and relevance settings, allowing results to be tailored to airport-specific use cases.

The team first introduced several practical changes. It built synonym mappings for airport and airline names and codes, then expanded those lists as new search patterns emerged. It also added fuzzy matching to account for spelling variations and typos, alongside auto-complete prompts designed to reduce input errors and guide users towards terms more likely to produce relevant results.

Discovery shift

Changi also added recommendation features to keep users engaged when a direct match is weak or unavailable. A "You May Also Like" section presents related options instead of letting a search end with a poor result or no result at all.

This approach draws on hybrid search ideas discussed in the OpenSearch community. In practice, the system can return both direct matches and alternative suggestions that may better reflect a traveller's intent.

Another notable change was the inclusion of iShopChangi in the same search environment. Product listings from the online duty-free and retail platform now appear alongside tenants, services and other airport information, creating a single entry point for both discovery and purchasing.

For an airport operator with a significant commercial footprint, that integration links wayfinding and service search with retail transactions. A traveller searching for a category, service or outlet can now encounter products available to buy online without moving to a separate platform.

Measured results

Changi said search relevance has improved markedly since the rollout. Zero-result queries have fallen from 17% to below 3%, suggesting fewer users are reaching dead ends when using the airport's digital tools.

According to the group, the "You May Also Like" feature recorded a click-through rate of about 11%. That suggests some users are engaging with suggested alternatives when the first result set does not fully answer their query.

The organisation also said searches that led to iShopChangi conversions accounted for a measurable share of overall sales on the eCommerce platform. While it did not disclose the value of those transactions, the figure suggests search is becoming more central to digital retail revenue.

The project reflects a wider pattern among large transport hubs and consumer-facing operators, many of which are revisiting search as a front-end service rather than a back-office function. In environments where users seek a mix of navigation, service information and products, search quality can influence both customer satisfaction and commercial performance.

Airports present an especially complex case because terminology is inconsistent and often highly contextual. Travellers may search by airline code, terminal name, brand, product type, common phrase or misspelt term, requiring systems to interpret intent rather than simply match exact labels.

By giving its team more direct oversight of search logic, Changi has sought to make that interpretation more responsive. The platform also provides a foundation for further work on intent-aware search, especially where product-level or tenant data may be incomplete.

Those efforts remain under exploration, but the current deployment already shows the operational effect of tuning search around specific user behaviour. Zero-result queries have fallen from 17% to below 3%.