How Virgin Australia is using machine learning to reach new travellers

Emily Bowers / July 2019

As more travellers kick off their research and planning by searching online, savvy brands are finding efficient ways to be present and relevant from the start. Read how Virgin Australia used a new automated solution in Search Ads 360 to streamline its search strategy and double bookings from non-branded queries.

Thanks to online travel aggregators and easy-to-access information, most Aussies have become their own travel agents. They regularly use multiple devices and sources to plan and perfect their trips, and they’re often more concerned with finding the best value than staying loyal to a particular brand. While this behaviour has created an even more complex consumer journey, it also gives brands countless opportunities to be considered when travellers go online to plan a trip.

In the last year, we’ve seen more and more travellers use search to start their planning. For airlines like Virgin Australia, showing up when consumers search for relevant information is no easy task — especially on generic, non-branded queries. Identifying and investing in the right keywords is a time-consuming process that requires constant upkeep, but it’s essential to connect with travellers when they’re curious, intent-driven, and open to new brands.

To continue driving sales in such a competitive market, Virgin Australia realised its search strategy needed a shake-up. So, the brand worked with the Google Marketing Platform team to test a new data-driven attribution (DDA) model and see whether automation in Search Ads 360 could help it capture attention across different devices and drive incremental sales from non-branded searches.

Bringing machine learning into the mix

Before testing an automated solution, Virgin Australia relied on a combination of linear and last-click attribution — either giving equal credit to each touchpoint that led to a conversion or 100% credit to the last click. But over time, the brand realised it didn’t have a clear enough view of which generic keywords were driving the most valuable conversions.

Automated channel grouping in Search Ads 360 helped Virgin Australia turn an estimated five-hour job into a five-minute job.

Virgin Australia decided to set up a different attribution model and test a new automated channel grouping feature in Search Ads 360 that uses machine learning to identify, segment, and label keywords based on their role in driving conversions. After training its new DDA model, the brand saw it could potentially optimise to 35% more revenue from non-brand search. Moreover, automated channel grouping in Search Ads 360 helped Virgin Australia turn an estimated five-hour job into a five-minute job. 

The results inspired the brand to make a full switch to DDA. The brand also used a target ROAS bidding strategy in Search Ads 360 with cross-device attribution to ensure it was making the most of its spend at every touchpoint, whether travellers were researching on desktop, tablets, or mobile.

Boosting revenue and bookings from non-brand search

Comparing its year-over-year results across non-brand search, Virgin Australia saw a 124% increase in revenue, a 139% increase in flight bookings, and a 40% increase in website traffic — all while maintaining return on ad spend and hitting the business’s KPI goals. The shift to DDA not only made a huge impact on the brand’s sales, but it also freed up time so it could reinvest in areas like creative strategy and optimisation.

DDA results from non-brand search


“Anything that helps us stay agile and focus on the big picture is worth testing,” said Ben Will, Virgin Australia’s digital marketing specialist. “Automated solutions in Search Ads 360 are sure to be a central part of our search strategy moving forward.”

From insights to intent: Why Search is more valuable than ever in Australia