Structured listing experience

Context

Obstacles related to seller knowledge, time limitations, and exhaustion have impeded the expansion of our customer-to-customer (C2C) base. A key factor contributing to reduced completion and conversion rates is insufficient listing quality. C2C sellers often overlook crucial item details, resulting in their listings going unnoticed by potential buyers who utilize specific filters during their search. In contrast, traditional search queries yield uniform outcomes for all listings, presenting buyers with a broad recall set.

The precision of search results is directly dependent on the precision of the initial query.

Filters you commonly see on any search result page are meant to augment the specifics that your search query may be missing. They simply offer a structured-data “shortcut” to getting to the specific (in our case) subset of items you’re searching for.

The problem

Buyers use item specifics to filter their search results, and your item will only appear in those filtered search results if you've added the matching item specific. The more item specifics you complete, the better eBay can match your item to what a buyer is looking for.

The problems begin when a seller neglects to provide item specifics in during the listing process. In the pre-AI era, eBay’s search algorithm simply fails to identify item’s specifics based on other cumulative attributes. And as a result, listings with missing specifics are excluded from search results, even for items that meet the shopping intent of the would-be-buyer.

So why don’t sellers provide listing specifics?

Sellers perceive this as extra work that is not tangibly justified. In many other cases they either do not know the specifics or where to look for them.

Why does eBay not make them mandatory?

This would dramatically affect listing completion rates. The downstream effect would be a dramatic reduction in GMV, conversions rate and revenue.

The solution

The only reasonable and sensible way forward would be to make it exceedingly easy and intuitive for sellers to find and provide the complete set of item specifics during listing flow.

The concept

A system of intelligent item specifics that would progressively adapt to user selections. Each subsequent subset would take into account the combination of already selected specifics.

For example, Brand will inform device type. A device type will in turn inform model, color and capacity, among other details. Tap-based specifics selection performed dramatically better in early studies where sellers had found the advantage not just in interactive simplicity but in a more accurate resulting listing.

Challenges

eBay is a vast marketplace with over 20,000 product categories. Any concept entering entering into consideration must be screened for scalability across verticals and major L1 - L2 categories. Once outliers are understood, next we must consider variations to the experience, as the effectiveness of the new tool will vary greatly. And while properly executed variations are a norm, it is critical to ensure that the tool will have coverage without major breaks.

Antiques and Collectibles had always presented a special challenge, as it is simply impossible to structure product data for centuries of craft with its million variations and styles.

Musical Instruments categories presented their unique requirements with aspect-sensitive listing process.

Often the specifics could not be tightened enough for eBay to set accurate pricing expectations as well as listing recommendations. I believe such outliers are permissible because they are inline with their specific vertical realities. But they nonetheless serve their purpose well in stress-testing concept’s integrity.

The need for specifics was greatest with Motors where accurate fitment was detrimental to the shopping success.

For some common electronics, the list ran deep into 25+ specifics, with each of them able to skew the entire product mapping entirely. For “hard” goods, such as these, it made sense to display a product card once we have reached the right point of informational intake. This would save seller time and effort for selecting the remaining specifics and finalize item’s mapping to a catalogued product.

Concept wireframe

This work resulted in a well-rounded and scalable model for quick and accurate listing experience. Key components: item specifics taped selection, Pricing recommendation and Shipping guidance.

8 years later

When you list an item on eBay today, you will benefit from the visionary bits born in 2016. Today’s listing experience runs on key functionalities such as dynamic aspects, product card (catalogue mapping) and prefilled specifics.

Dynamic aspects

Dynamic aspects guidance ensures an uninterrupted progressive aspects collection. All the while your listing is already being created in the background.

Product card

After gathering initial aspects, the threshold of confidence is reached and eBay will surface catalogued products that may be a 100% match to your item. If selected, your listing will created with one click, including all item specifics.

Prefilled specifics

With a single click eBay seller ensures that their listing will be visible to all buyers who are shopping for their item, exponentially increasing the likelihood of a match and a resulting conversion.

Next: eBay Valet

Amplifying eBay’s role in solving for household clutter problems while promoting charitable causes and reinforcing green recycling practices.