Experiment: How Careem lifted conversions with automated display ads

November 2019
Experiments banner.png

What we set out to test

Can brands increase conversions by incorporating machine learning across display campaigns?

The background

Careem is one of the top transport network companies in the Middle East, Pakistan, and North Africa. The brand offers ride-hailing services in more than 90 cities across 14 countries, so one of its top priorities is encouraging new drivers to join its team.

Although Careem primarily uses print flyers to recruit new drivers, it recently started engaging prospective leads online with search and manual display ads. To take its digital efforts one step further, the brand decided to test automation across its display campaigns.

How we set the experiment up

The brand created Smart Display campaigns to automatically bid on keywords, optimize ad placement, and develop more relevant ads. Careem then set up a pre/post test in Tier 1 cities in Pakistan to see how machine learning influenced its display campaign’s performance.

  • Pre-automation: Search ads and manual display ads for eight weeks
  • Post-automation: Search ads and Smart Display ads for eight weeks

The brand considered each sign-up for its recruitment events as a conversion. To uncover the best way of boosting incremental conversions, Careem compared two things at the end of its experiment:

  • Online and offline campaigns
  • Manual and automated display campaigns

Solutions we used

What we learned

Creating, placing, and delivering ads with Smart Display campaigns can help brands boost conversions and reach relevant audiences at scale. Careem learned that it could engage and recruit new drivers more efficiently by fueling its digital campaigns with automation.

The brand saw a 320% increase in sign-ups with its Smart Display Campaigns than it did with manual display ads. By automating its display campaigns, the brand was able to decrease cost per acquisition (CPA) by 27% when compared to offline recruitment efforts.


“Our experiment showed us how important it is to continuously test new technologies in our campaigns. We’re now exploring even more ways to grow our team with machine learning.” — Michael Ruder, Senior Digital Marketing Manager, Careem

This case study is part of the Experiment with Google Ads program.

Experiment: How Uniqlo drove in-store sales by connecting with high-value app users