{"id":20728,"date":"2022-03-18T09:47:21","date_gmt":"2022-03-18T06:47:21","guid":{"rendered":"https:\/\/quicklizard.wpenginepowered.com\/blog\/online-travel-agency-case-study\/"},"modified":"2026-09-30T08:48:05","modified_gmt":"2026-09-30T05:48:05","slug":"fallstudie-uber-ein-online-reiseburo","status":"publish","type":"post","link":"https:\/\/quicklizard.com\/de\/blog\/online-travel-agency-case-study\/","title":{"rendered":"Online-Reiseb\u00fcro Fallstudie"},"content":{"rendered":"

The Client<\/h2>\n<\/div>\n\n\n

Ein in Europa ans\u00e4ssiges Online-Reiseb\u00fcro. 70% der Verk\u00e4ufe werden \u00fcber Metasuchmaschinen get\u00e4tigt. Die verbleibenden 30% des Umsatzes stammen aus direktem Verkehr, der sich aus organischem Verkehr, Keyword-Werbung und SEO zusammensetzt.<\/p>\n\n\n

The Main Challenges<\/h2>\n<\/div>\n\n\n

Online travel agencies are at the forefront of dynamic pricing. They need to utilize every piece of information they have to increase their profit margins while staying competitive.<\/p>\n\n\n\n

Different tactics may be used for different product types (flights, hotels, packages, cars, cruises) and channels (organic, SEO, SEM, social, meta search).<\/p>\n\n\n

\n
\n

The booking engine<\/p>

The customer developed an advanced booking engine that combines results from multiple GDSs, low cost airlines and charter flight options to provide travelers with the lowest-priced options on the market.<\/p>\n<\/div> <\/article>\n

\n

Before Quicklizard<\/p>

However, their home-grown system required pricing managers to manually determine their commission markup from net cost. The pricing managers didn\u2019t consider competitor pricing or environmental factors like time of day or weather into their pricing, and applied changes rarely, due to limitations of manual work.<\/p>\n<\/div> <\/article>\n <\/div>\n<\/div>\n\n

The QL Solution<\/h2>\n<\/div>\n\n
    \n
  1. \n

    Conversion prediction and price simulation<\/p>

    QuickLizard implemented a hybrid machine learning and a\/b testing approach. The platform predicted the probability of conversion for each customer based on user behavior and competitor pricing. A simulator then recommended the optimal price point for each permutation.<\/p>\n<\/div> <\/li>\n

  2. \n

    An accurate cost function<\/p>

    Additionally, an accurate cost function was applied to correctly reflect parameters not included in the net fare, including ancillary booking, quota-based incentives from vendors, and halo-effect sales.<\/p>\n<\/div> <\/li>\n

  3. \n

    A bidder for meta search<\/p>

    The recommended markup was applied by pricing managers using the existing home-grown system. The meta search traffic utilizes a separate pricing tool. QL implemented a bidder that uses machine learning and reinforcement learning to predict the lowest bid offered by the competition for each itinerary. Our client used this information to create a set of rules and constraints to propose a price.<\/p>\n<\/div> <\/li>\n <\/ol>\n<\/div>\n\n

    Das Ergebnis<\/h2>\n<\/div>\n\n