Bots, scraping and look-to-book on travel websites

No longer just a security problem

Automated traffic has always existed in travel, but its nature has changed. Price scrapers and bots that hold seats without paying — so-called seat spinning — are now joined by AI agents querying availability on behalf of users or platforms. The World Aviation Festival 2026 agenda tackles it from two angles: how bots and agents threaten revenue and customer experience, and how to use AI to solve the look-to-book problem that AI itself has made worse.

For ecommerce, the impact is not only technical. It affects measurement, costs and the availability real customers see.

Four effects on the business

Distorted conversion

If automated sessions are not excluded, conversion falls without anything changing for customers. Product decisions and A/B tests are made on a contaminated baseline. That is why it is key to calculate travel conversion correctly.

Infrastructure and distribution cost

Every search consumes own capacity and often queries systems that charge per transaction. An out-of-control look-to-book ratio can turn the website into a cost centre.

Blocked inventory

Bots that start bookings and never pay hold seats or rooms that cannot be sold to real customers, sometimes on the highest-demand dates.

An edge for competitors and intermediaries

Systematic scraping lets third parties replicate prices and react first, reducing the value of direct-channel exclusive fares.

Separating good bots, bad bots and legitimate agents

Not all automated traffic should be blocked. Search and metasearch crawlers bring demand; some AI agents will act on behalf of real customers. The right strategy is to classify and govern:

• Necessary bots: search engines, metasearch and contracted partners; allowed within limits.

• Legitimate agents: routed to authenticated APIs with their own terms, not to the website.

• Abusive bots: mass scraping, inventory holding and fraud; blocked or disincentivised.

Practical measures

1. Measure before blocking. Quantify how much traffic and search volume is automated and where it comes from.

2. Clean the analytics. Exclude non-human traffic from funnel reports and from the events measured in the booking funnel.

3. Protect the points of value. Availability search, inventory holds and checkout, with controls proportional to risk.

4. Offer an official door. APIs for partners and agents, so legitimate access does not need to impersonate a human.

5. Review hold rules. Inventory hold times and limits per session or identity.

Server-side measurement helps distinguish real from automated traffic, as explained in server-side tracking and ecommerce measurement.

Indicators to track

• Share of automated traffic by type and source.

• Human versus total look-to-book ratio.

• Cost per search and per booking.

• Bookings started without payment on high-demand dates.

• Clean human conversion, the baseline for any booking engine optimisation.

Conclusion

Automated traffic has moved from noise to a factor that shifts revenue, costs and metrics. Businesses that measure it, classify it and offer official channels to legitimate agents will protect their inventory and make decisions on reliable data.

Consumer Services Hub - Strategic ecommerce consultancy for B2C service companies

consumerserviceshub.com