Recommending What to Search: How Query Suggestions Drive Sales and Diversity in Mobile Shopping

Why It Matters
As mobile commerce continues to dominate online activity, platforms must help users search more effectively on small screens. Our research shows that query recommendations — automated suggestions of what to search — can increase purchases and broaden consumers’ exploration, benefitting both users and businesses.

Key Takeaways
• A query recommender system increases sales volume by 1–2% and promotes exploration across more products and sellers.
• It fosters more diverse consumer choices, reducing over-reliance on popular sellers and promoting new merchant discovery.
• When paired with auto-completion, query recommendations make search easier and more engaging for users.

 

How Do Query Suggestions Influence Mobile Shoppers?

This study investigates how a Query Recommender System (QRS) — which suggests search terms to users — affects consumer behaviour on a large mobile food delivery platform in Asia. While most prior research has focused on recommending products, our study focuses on recommending queries, a tool that shapes consumer behaviour earlier in the decision process.

We conducted a large-scale randomised field experiment on over 58,000 consumers. Treated users received AI-powered query suggestions in the search bar for 30 days, while control users only saw the usual auto-completion feature.

The results were clear: treated users tapped on more merchants, placed more orders, and explored more diverse categories than users in the control group.

Expanding User Choices While Easing the Search Process

The QRS encouraged users to type shorter and more generic queries. This reduced cognitive effort and encouraged exploration beyond their usual purchases. Users were more likely to interact with new merchants and less likely to repeat past purchases. We observed higher individual and market-level diversity in purchase behaviour.

Importantly, query suggestions and auto-completion worked hand-in-hand: users in the treatment group used auto-completion more and manual typing less. This complementary relationship made it easier for users to search efficiently and broadly at the same time.

Creating a Healthier Digital Marketplace

The combination of increased purchase volume and more balanced consumption patterns means QRS benefits not just users, but also new or smaller sellers who often struggle to compete for visibility.

By fostering greater exploration and lowering search friction, query recommenders help level the playing field. Treated users showed more equitable engagement with a wider range of sellers, a trend confirmed through Gini coefficient and Lorenz curve analysis.

The platform eventually rolled out the query recommender system to all users — recognising its business value in enhancing both sales and customer satisfaction.

Business Implications
Our study offers concrete evidence that investing in query recommendation tools can yield measurable benefits. Platform operators can:

• Improve user engagement and conversion through reduced search friction.
• Increase overall sales volume while reducing market concentration.
• Encourage exploration of lesser-known products and merchants, supporting ecosystem sustainability.

Firms should view QRS and auto-completion not as substitutes, but as complementary tools that together improve search experience and market outcomes.

 

Authors & Sources
Authors: Shuang Zheng (Dalian University of Technology), Siliang (Jack) Tong (Nanyang Technological University), Hyeokkoo Eric Kwon (Nanyang Technological University), Gordon Burtch (Boston University), Xianneng Li (Dalian University of Technology)

Original Article: Marketing Science

---

For more research, click here to return to NBS Knowledge Lab.