Look-to-book ratios: why travel search costs money and how to control it

The look-to-book ratio is the number of searches a seller sends to a supplier for every booking it completes. A platform sending 2,000 availability requests per booking has a ratio of 2,000:1. Suppliers track it because searches consume their capacity, and many contracts set limits or charges when ratios exceed agreed levels. For a travel platform, it is one of the few technical metrics with a direct line to the profit and loss statement.

Why it matters commercially

Search traffic grows faster than bookings. Metasearch, price comparison, bots and repeat browsing all increase searches without increasing sales. If supplier contracts include ratio limits or excess-search fees, a successful marketing campaign that drives browsing can turn into a cost problem. Some suppliers throttle or suspend access when ratios stay high.

Look-to-book is therefore not only an engineering concern. Product, commercial and engineering teams should share one dashboard that shows it per supplier.

Where excess searches come from

• Bots and scrapers collecting prices from your site.
• Metasearch traffic that queries frequently but converts at a lower rate.
• Duplicate requests from retries, double clicks or parallel UI components.
• Over-wide fan-out — querying every supplier for every route, even when some never have relevant content.
• Flexible-date and calendar features that multiply searches per user action.

Engineering levers

Cache with intent

Caching is the strongest lever, as long as it respects price volatility. The approach in caching flight search — layered caches, variable TTLs and mandatory re-pricing — reduces supplier calls without selling prices you cannot honour.

Deduplicate in flight

If two identical searches arrive within seconds, send one supplier request and share the result. A short-lived lock keyed on the normalised search, as described in Redis beyond caching, handles this cleanly.

Route-aware fan-out

Not every supplier serves every market. Maintain a routing table that maps origin, destination and product type to the suppliers likely to have content, and update it from booking data. Fewer calls per search lowers both cost and latency.

Protect against automated traffic

Rate-limit by client, detect bot patterns, and serve cached or reduced results to suspicious traffic. For partners with API access, set explicit quotas per key.

Budget expensive features

Calendar views, flexible dates and “nearby airports” are valuable but costly. Serve them from cache or pre-computed data rather than live fan-out.

Improve the other side of the ratio

A ratio has a denominator. Better conversion lowers look-to-book as effectively as fewer searches: fast pages, accurate prices, clear fare brands and a checkout that handles price changes gracefully all turn searches into bookings.

Measure it properly

Track per supplier and per channel:

• searches, price checks and bookings per day,
• the ratio over rolling windows, against contracted limits,
• cache hit rate and the price-change rate at re-price,
• top sources of traffic by client or partner.

Alert before limits are reached, not after a supplier sends an invoice.

The takeaway

Look-to-book turns search architecture into a commercial question. Platforms that cache carefully, route searches intelligently, filter bots and keep improving conversion can grow traffic without growing supplier costs at the same rate.