Meridian Journeys
Meridian ran TravelSpy across 340 departures for a full season. Demand scoring identified weak departures in week four rather than week twenty, and targeted packaging replaced blanket discounting entirely on two thirds of the programme.
booking conversion
Measured against a randomised holdout
Northwind Air
By resolving anonymous fare searches into modelled travellers in-session, Northwind reordered ancillary presentation per traveller and began packaging accommodation at fare selection. Trip margin previously captured downstream stayed in channel.
ancillary attach
Measured against a randomised holdout
Casa Lumera
Casa Lumera used direct-intent detection to identify guests mid-comparison and responded with a packaged value add rather than a rate cut. Commission costs fell while average guest spend rose.
direct booking share
Measured against a randomised holdout
Aurora Voyages
Long research cycles were finally visible. Aurora surfaced a written dossier in the agent panel before connect, cutting handle time while lifting qualified enquiry volume by forty-four per cent.
average handle time
Measured against a randomised holdout
Saltwater Collective
Modelled value per visitor fed directly into bidding. Placements that produced researchers rather than bookers were defunded within two weeks, and budget shifted to sources with proven booking value.
cost per booking
Measured against a randomised holdout
Highfield Travel
Automatic first-draft itineraries removed the speculative build. Advisors moved from logistics assembly to relationship selling and the agency grew revenue without growing headcount.
itineraries per advisor
Measured against a randomised holdout
Because a conversion claim without a method is just marketing.
A fixed percentage of eligible travellers are excluded from all TravelSpy activation for the full programme, not just a test window.
The primary metric and the measurement period are agreed before the programme starts and cannot be changed retrospectively.
Results are reconciled against your own reporting stack. Where they disagree, your numbers are the ones we publish.
Book a 30-minute session. We will run your own traffic through the model and show you the segments, the predicted spend bands and the packages you are not selling yet.