Answers · OTAs & metasearch · 8 questions
Updated 25 Oct 2026 · Reviewed by Mohammad Shakir, Co-founder & CEO
Once price intelligence is a system rather than a spreadsheet, the questions become architectural: API or feed, how many checks, what format, what it costs at scale. These answers cover the data plumbing behind a pricing operation.
An API answers a request for a specific hotel and date on demand. A feed delivers a scheduled batch of data. APIs suit live use; feeds suit analytics.
Many teams use both: an API for pricing engines and a feed for reporting.
It depends on inventory, dates covered and refresh rate. Multiply properties by dates by sources by refreshes per day, then prioritise by demand to keep it affordable.
A tiered approach works: frequent checks for high-traffic properties and near dates, lighter checks elsewhere. Estimate before you buy, since per-call pricing scales linearly.
Prioritise by demand, cache what does not change quickly, avoid duplicate checks and use cheapest-room checks when full detail is unnecessary.
Review which checks actually change a decision and drop the rest.
JSON via API for engines, CSV or warehouse tables for analysts, and a dashboard for business users. Pick the one your team already works in.
Stable field names and versioning matter more than the format.
Share the source and a sample of target properties. The provider builds and validates a collector, then it runs alongside existing sources.
Ask for the typical turnaround and whether adding a source changes your pricing.
Flights are tracked by route, date, cabin and fare rules, and change faster. Hotels add room type and cancellation complexity.
Both need normalisation so that two prices are actually comparable.
Break the package into components where possible, then compare the total payable price for identical inclusions and dates.
Inclusions differ widely between packages, so a matching step matters.
Common models are per call with volume slabs, or a platform fee with included volume. Per-call pricing aligns cost with usage.
Check what counts as a billable call, and whether failed calls are billed.
ScrapeGuys builds and operates production scraping APIs that return clean, structured JSON - and takes custom targets live in 48-72 hours.