Of your peers have already watched this video.

2:30 Minutes

The most insightful time you'll spend today!

Case Study

How e-Com Firm Bukalapak Achieved 5X ROAS With Machine Learning

In 2018, Indonesia accounted for 94% of SEA’s $23 billion e-commerce industry. Today, the country’s massive e-commerce sector continues to grow, along with the number of brands looking for innovative ways to compete for a piece of the pie.

As one of the largest e-commerce companies in the regionBukalapak receives a high volume of website visitors via direct traffic, Shopping ads, Google Display Network ads, and YouTube ads.

But when the brand noticed too many potential customers were browsing its website without converting, it knew it had to reconsider its marketing strategy. In an effort to reach consumers who were more likely to buy its products, Bukalapak turned to Smart Shopping campaigns.

Experimenting with Automation

By combining standard shopping and dynamic remarketing campaigns, Smart Shopping campaigns use automated bidding and ad placement to promote products to users across Search, Display, and YouTube. The automated solution also allows brands to reach high-value users who have already seen its ads or visited its site directly but left without converting.

Always open to trying new strategies, Bukalapak launched a three-month Smart Shopping campaign focused on 5% of its Shopping ads traffic using a maximize conversion value bidding strategy. The rest of the brand’s traffic (95%) was assigned standard shopping and dynamic remarketing campaigns, and the results of these were measured against the automated alternative.

The team was able to launch the Smart Shopping campaign with little manual effort by:

  • creating a separate campaign with a determined traffic split and a recommended daily budget.
  • uploading the Bukalapak logo and image banner for responsive display ads.
  • designating Indonesia as the country of sale.

The campaign combined the brand’s existing product feed with Google’s machine learning algorithm to serve more than 40 million products to potential customers across multiple channels — all while automating ad placement and bidding for maximum conversion value.

Smart Shopping Campaign Saves Time, Boosts ROAS

The Smart Shopping campaign achieved 5X higher ROAS than the standard shopping effort while also driving 4X more conversions and 300% growth in conversion value, leading to 2.5X more new customers.

“The automation not only allowed the team to focus less on manual campaign optimization but also helped them boost relevance among high-value users,” said Tushar Bhatia, associate vice president of growth at Bukalapak.


The impressive results encouraged Bukalapak to increase its investment in Smart Shopping campaigns by 27X over the past year. The brand plans to remain at the forefront of innovation by continually testing new products and further optimizing its campaign strategies.

More Relevant Stories for Your Company

Case Study

How Google Helped the Indian Govt Choose Airport Sites: The Inside Story

Setting out a plan for facility locations is a classic challenge for organizations of all shapes and sizes. This is the case for companies around the world—from retailers to governments and corner stores to department stores. But, in India, it's an especially hard problem to crack. Healthy returns and high


How AI Can Predict Consumer Demand and Retain Customers

How can AI help you target customers who are looking for products like yours? The challenge: This customer wants to buy the best cat food for her pets. Reach the Right Customers How can AI help you target customers who are looking for products like yours?The challenge: This customer wants


Engage and Translate: Text and Audio Chat in 100+ Languages

As users worldwide connect to the internet and work remotely, it’s more important than ever to make interactive applications chat in many languages. But when users speak 100s of languages, this quickly can get challenging. Where to start? If you are new to translation, Sarah Weldon, the Product Manager for


What are TPUs and Why Should Data Scientists Care?

When it comes to machine learning, more data means better results. But processing more data also requires more computing power. CPUs are great for sequential arithmetic calculations. They offer low latency with this type of workload. But to speed up machine learning, models have to perform multiple calculations in parallel.