What we set out to test
Can brands encourage high-value consumers to make in-store purchases by fueling their campaigns with machine learning?
Uniqlo is the world’s third-largest apparel brand, and its goal is to become even more customer-centric by continuing its digital transformation. Although e-commerce accounts for 14.6% of all retail spend in APAC, the brand knew its digital experiences had to extend to brick-and mortar-stores because many people still prefer to try on items before checking out.1
To make sure its digital experiences were helpful to in-store shoppers, Uniqlo developed an in-app barcode that people could scan at checkout to take advantage of exclusive promotions. Although the brand already had a large app user base, it wanted to find a new way of connecting with high-value consumers and encouraging them to shop in stores.
The brand had previously used App campaigns for installs (AC for installs) to drive downloads. Intrigued by machine learning’s potential, Uniqlo set out to discover whether it could use automation to boost in-store purchases rather than downloads.
How we set the experiment up
Uniqlo set up a controlled pre/post-test in Japan to see if it could entice more shoppers to make purchases at brick-and-mortar stores by tapping into machine learning.
- Pre-test: AC for installs for two weeks
- Post-test: App campaign for action (AC for action) for two weeks
Throughout its experiment, Uniqlo kept its point of sales and ad distribution data updated within a seven-day window to make sure machine learning could create and place ads in a way that was tailored to app users. After four weeks of testing, Uniqlo compared the two campaigns’ respective impact on in-store purchases.
Solutions we used
What we learned
Machine learning can help brands encourage high-value consumers to take a specific action by delivering the right ads to the right audience. Uniqlo found that AC for action drove 3X more in-store purchases than AC for installs. And by integrating its data with its ACs, Uniqlo could clearly visualize how its ads influenced app users’ shopping decisions. Moving forward, the brand is excited to use machine learning to promote its new StyleHint app that offers shoppers fashion inspiration.
This case study is part of the Experiment with Google Ads program.