FIELD GUIDE 01
From Global API to a Real Booking Service
Supplier connectivity is the beginning of a booking product, not its finish line.

A supplier response can be technically correct and still be unsuitable for a customer decision. The work begins when an API payload has to become a trustworthy product, an explainable booking promise and a record that an operations team can use later.
FIELD NOTES
Why this matters
Travel suppliers optimise an API for their own inventory, rules and operational boundary. A customer-facing service has a different responsibility: it must present what a traveller can choose, what they are agreeing to and what may still change.
That distinction is easy to miss when an integration first returns data. Availability may be stale, a price may be conditional and a product label may assume knowledge that a local customer does not have.
Where the real problems appear
Authentication, rate limits, product metadata, price, availability, booking state and cancellation rules each arrive with different reliability and timing. A response called available is not always a promise that an order can be confirmed.
Pending, confirmed, failed and cancelled are examples of service states, not a universal SORADA standard. The important decision is to define what each supplier response means for the customer and for the next operational action.
The supplier model and the service model
Supplier fields are not a customer information architecture. A terse fare code may hide restrictions; a price may be quoted at a particular moment; an error may be meaningful to an engineer but alarming or useless to a traveller.
A service model creates explicit product meaning around those fields. It decides what is comparable, what needs explanation and what must be held back until a later confirmation.
Booking journey: API and product together
Search establishes a useful set of possibilities. Product and schedule pages turn source data into a choice; seat or option selection exposes constraints; price and confirmation explain the commitment before payment.
After payment, the result cannot be treated as a decorative success screen. It must communicate the current booking state, the supplier reference where appropriate, delivery expectations and a safe route when the response is pending or cannot be completed.
Technical view
Keep source responses, normalised service data and customer-facing copy distinct enough to be inspected. Rate-limit handling and retry behaviour should not create duplicate customer actions. Error handling needs a controlled path rather than leaking raw supplier messages.
Time-sensitive fields deserve their own rules. A product cache may be useful for browsing while price or booking confirmation requires a more current check.
Product and UX view
A booking interface should state conditions at the point where they change a decision. Customers should not need to infer whether a schedule is local time, whether a fare is flexible or whether a selected option is compatible with the product.
The aim is not to expose every upstream field. It is to create enough context for an informed commitment and an understandable result.
Operations view
After booking, change, cancellation, status checks, customer questions and supplier rechecks enter the picture. The operating team needs a path from the customer record to the source state without rebuilding the journey from scattered screens.
A clear source reference, a recorded customer-facing result and a traceable history make that work calmer when an expected response changes.
SORADA viewpoint
Global API work becomes product engineering when the team owns the gap between a supplier rule and a customer promise. That is where data architecture, booking UX, localization and operations meet.
PRACTICAL TAKEAWAYS
Questions to carry into the next design.
- Map supplier states to explicit service meaning.
- Design the booking result as an operational record.
- Use product language to explain conditions before payment.
CHECKPOINTS
- What does each availability and booking state mean to a customer?
- Which price, rule and schedule values must be checked again before commitment?
- Can a raw supplier error be translated into a safe next action?
- Can an operator see what the customer saw at the time of booking?
COMMON MISSTEPS
- Treating an availability response as a confirmed reservation.
- Showing supplier terminology without the conditions needed to interpret it.
- Designing only the happy-path confirmation and adding support later.


