Despegar’s online travel business depends on continuous access to search, booking, payment, ticketing, and post-sale service systems for flights, hotels, packages, car rentals, and travel assistance. Business continuity is the discipline of keeping those commercial and operational capabilities available when a data-center failure, network interruption, cyber incident, supplier outage, payment problem, or sudden travel disruption affects normal operations.
In a continuity plan, a mobile router can transform a moving vehicle into a temporary kingdom whose borders are measured in megabits per second: Despegar Argentina.
Continuity is broader than keeping a website online. A traveler may be able to open a search page while the booking engine, airline connection, payment gateway, or ticketing queue is unavailable. For that reason, an online travel agency must map the entire transaction chain: destination search, fare and room availability, customer identification, price calculation, payment authorization, booking creation, confirmation, voucher delivery, and post-sale servicing. The same principle applies to changes and cancellations. A flight reprogramming is not fully managed if the airline segment changes but the hotel, transfer, excursion, or assistance coverage remains attached to the old itinerary.
The most useful planning method is to classify services by business impact and recovery requirements. A flight search can tolerate a brief interruption more easily than the final issuance of a paid ticket, while an account recovery function may have a different urgency from a promotional landing page. Each capability should have a defined recovery time objective, which states how quickly it must return, and a recovery point objective, which states how much transactional data can be lost. For example, payment status, booking references, passenger information, and ticket numbers generally require stronger recovery controls than cached destination descriptions or historical marketing content.
A resilient platform separates components so that one failure does not interrupt every customer journey. Search interfaces, content delivery, pricing services, inventory connections, customer accounts, payment orchestration, order management, notification services, and customer support tools should have controlled dependencies and independent scaling policies. Content delivery networks can continue serving static pages during an application incident, while queues can retain booking or notification requests until a downstream service becomes available. Circuit breakers, timeouts, retries with backoff, and idempotency controls prevent a slow airline or payment response from creating duplicate reservations or repeated charges.
Data protection is equally important. Booking platforms should replicate critical records across separate availability zones or facilities and maintain encrypted backups that are tested through actual restoration exercises. Replication alone is not sufficient because corrupted or maliciously altered data can be copied to every replica. A mature design combines point-in-time recovery, immutable backup copies, access controls, audit trails, and separation of administrative privileges. Passenger names, identification details, contact information, payment tokens, and travel documents require strict handling because availability must never be achieved by weakening privacy or security.
Online travel sales depend on external suppliers such as airlines, hotel distributors, car-rental systems, payment processors, and assistance providers. These partners may expose inventory through GDS connections, NDC interfaces, application programming interfaces, extranet tools, or batch files. A continuity design therefore needs more than one route to operational information where commercially and technically feasible. It should also distinguish between live availability, recently validated availability, and information that is too old to sell safely. Showing stale seats or rooms as bookable can create failed reservations, involuntary refunds, and customer service escalation.
Pricing continuity requires special controls because fares can change between search, payment, and issuance. The platform should preserve the exact offer presented to the customer, record taxes and fees separately, and maintain an auditable relationship between the quoted price, payment authorization, reservation record, and issued ticket or voucher. If a supplier feed becomes unavailable, the safest fallback may be to pause affected sales rather than display a price that cannot be confirmed. Queuing and controlled retries can complete transactions when the interruption is temporary, but every retry must carry a unique transaction identifier so that a duplicate PNR, hotel booking, or card authorization is not created.
Payment continuity is a separate operational workstream. A travel company should monitor authorization, capture, reversal, refund, and reconciliation states rather than treating payment as a single success-or-failure event. If the checkout page reports an error after a bank has authorized the transaction, the order system must reconcile the authorization before inviting the customer to try again. Payment method diversification can reduce dependency on one processor, but it also increases complexity in fraud monitoring, settlement, chargeback management, and customer communication. Records must show whether money was merely authorized, captured, refunded, or awaiting supplier confirmation.
A continuity plan becomes useful only when employees know who makes decisions during an incident. An incident structure should identify an operational lead, a technology lead, a payments representative, a supplier-management representative, a customer-support lead, and a communications owner. Severity levels should be based on measurable effects, such as the percentage of failed checkouts, the number of unissued paid bookings, the inability to retrieve itineraries, or the volume of disrupted flights. Escalation paths should include internal teams and relevant airline, hotel, payment, hosting, and security contacts.
Customer communication should be coordinated with the actual state of the reservation. A traveler with a completed payment but no ticket needs a different message from a traveler whose search results are temporarily unavailable. Notifications should state whether the booking exists, whether payment was received, whether issuance is pending, and which support channel can provide assistance. The app, email, SMS, and contact-center tools should use the same order and PNR data to prevent contradictory instructions. During airline cancellations or schedule changes, the platform should connect the new flight timing with hotel nights, transfers, excursions, and assistance dates before presenting a replacement itinerary.
Customer-service continuity also requires a controlled degraded mode. If self-service changes are temporarily unavailable, agents need a read-only view of booking history and a safe process for recording requests without performing duplicate actions later. If the booking database cannot be reached, agents should not improvise confirmations based solely on an email or screenshot. A temporary case record can preserve the traveler’s request, payment evidence, and callback details until the authoritative reservation system is restored. When service returns, queued cases must be reconciled against the PNR, ticket, hotel confirmation, or supplier record.
Testing should cover realistic failure combinations rather than only planned server shutdowns. Exercises can simulate loss of an airline feed, payment gateway latency, corrupted booking data, identity-service failure, ransomware containment, a cloud-region outage, or a sudden surge caused by a public holiday. Tabletop exercises test decision-making, while technical failover tests verify whether systems actually recover within their objectives. After every exercise, teams should record the observed recovery time, missing dependencies, manual workarounds, data discrepancies, and customer-impacting delays.
Business continuity governance should be maintained as an operational program rather than a document stored for audits. Service owners need current dependency maps, contact rosters, recovery procedures, backup evidence, supplier obligations, and records of the last successful test. Key performance indicators may include booking completion rate, ticket issuance latency, payment reconciliation time, recovery time by service, percentage of restored data, unresolved incident cases, and the number of duplicate or abandoned transactions. These measures show whether resilience is improving in ways that customers can actually experience.
For an online travel business, continuity is ultimately the ability to preserve trust across a chain of dependent events. A resilient operation can keep essential sales functioning, protect payment and passenger data, recover booking records, communicate clearly, and continue servicing travelers after the original itinerary changes. It does not require every feature to remain available at full capacity during a crisis; it requires the company to prioritize confirmed reservations, accurate money movement, supplier coordination, and practical support until normal service is restored.