bonprix has always been a pioneer in the digital retail space. As an early adopter, they launched their first Shopping campaign back in 2012 to support their Swiss expansion. Since then, they’ve always looked for ways to continue innovating and growing, so they were keen to try Google’s new machine learning capabilities to automate their bidding.
Until recently, they made daily bid optimisations for individual products to maximise their return on ad spend (ROAS). bonprix decided to test the Target ROAS Smart Bidding strategy to help achieve their ambitious targets and save time from manual optimisations. Using their own imported conversion data, the Target ROAS bid strategy would automatically make query-level bid adjustments to maximise ROAS based on user signals, such as location, time of day, device type, and search query.
Alongside this bid strategy, bonprix also moved away from a granular item ID campaign structure. Instead, they created a new structure based on ROAS targets using their own historical data, then segmented by product category for reporting. The campaign with a lower ROAS target would allow the bid strategy to have more flexibility in bidding higher to remain competitive for high-profile products, while a higher ROAS target would help generate revenue efficiently on their core products.