Ecommerce search and recommendation has recently received attention in the academic literature because it is the primary way users find products on eCommerce sites—organisations such as Amazon, eBay, and Airbnb are completely dependent on search to match users with products. But eCommerce search is unlike web search in many ways, and is increasingly requiring sites to adapt to, or completely change, the way they work. The community has been addressing these differences in workshops held at SIGIR in 2017, 2018, and 2019. And we are seeing an increasing interest in eCommerce search papers at Information Retrieval conferences, and in Information Retrieval journals. For this special issue, we distributed a call for papers on eCommerce Search and Recommendation and have chosen 2 papers from the 8 submitted—all reviewed by a committee of international experts on the topic of the paper.

In “Informational, Transactional, and Navigational Need of Information Relevance of Search Intention in Search Engine Advertising”, Schultz (2020) discusses a 4-year study into online advertising for products—from the perspective of the product seller. He divides behaviour into the intent categories of transactional (keywords include the company services), and navigational (keywords include the company brand), informational (include other keywords), and studies user behaviour with and without the organic search results. Most importantly, he examines how behaviour changes between intents and what leads to a sale.

In “Deep Cross-platform Product Matching in E-commerce”, Li et al. (2020) discuss cross-site product matching. They observe different sites describing the same product in different ways. Not only can titles differ, but so too can the structured data. A product might be described as having a size in millilitres on one site but a volume in fluid ounces on another. The users just want the cheapest vendor regardless of the description, and so a way to match products across multiple sites is needed. They devise a neural model to do this and test it under two scenarios: “are these two products the same”; and “rank these products on similarity to this query product”.