CASE STUDY: How Zuji boosted bookings with personalised display ads

Background

Many visitors to Zuji Hong Kong search for outbound flights from Hong Kong but don't complete the bookings. Meanwhile, Zuji was seeing stagnation in its ROI (return on investment) for performance-marketing channels. In particular, display was lagging behind channels such as search-engine marketing. Working with Sociomantic Labs and MEC, the brand sought to create a data-driven, personalised customer experience to bring visitors back to the site to complete their bookings.

Execution

Real-time bidding and dynamic messaging enabled Zuji to reach individual users as segments of one, with personalised messages based on their interests, preferences and online behaviours.

By plugging into a stream of Zuji data showing users’ searched-for destinations, departure dates and flight offers in real time, Sociomantic powered four campaigns: one per language (English and Chinese) per platform (display and Facebook Exchange).

Sociomantic’s proprietary real-time bidder analysed user profiles, along with performance and yield data, to identify users who had been on Zuji’s website and had shown interest in outbound flights from Hong Kong. Based on the data, which indicated each user’s interest and hence price elasticity for a given flight, the engine calculated a bid price for the ad impression, based on Zuji’s ROI goals.

Example dynamic display units

The objective was to increase the likelihood of conversion by targeting only the most relevant users with personalised ads served at an optimal price. In contrast to the approach of showing all potential customers a limited set of offers, the real-time approach delivered incremental revenue because Zuji paid for ad impressions according to each user’s value.

Whenever the real-time bidding engine won an impression, Sociomantic’s Dynamic Creative Optimization (DCO) engine generated and displayed personalised flight recommendations to the individual user based on the same data used to win the impression. Related routes were also included in the ad messages. The DCO engine creates ad banners in HTML5 format, rather than Flash, to ensure readability on any device.

The platform also protected the Zuji brand by capping ad frequency on a per-user and per-time basis, which also maximises budget efficiency by reaching a larger number of users with higher value.

Results

Zuji’s investments in the Sociomantic campaigns generated a 100-fold return on ad spend, according to the companies.

From January through June, the campaigns delivered a 14 per cent increase in digital bookings.

In addition, Sociomantic campaigns also helped Zuji upsell hotel and package bookings, which increased by 10 per cent.

The campaigns outperformed previous retargeting efforts by at least three times, according to the companies. Based on this success, Zuji is using the dynamic-messaging strategy for its Hotels section as well.