Jane Li

Who they are

The sources read for this page establish Jane Li as a researcher at Google at the time of the paper, since the paper is listed on Google's research publications site, and as the first-named author of the work on good abandonment. They do not give a biography, a career history or a current position, and this page does not supply one.

That limit is deliberate. The publication prefers a short page that states what is known to a longer one that fills the space with plausible detail. What can be said with confidence concerns the paper, so most of this page is about the paper.

Credit matters here for a practical reason. Papers with several authors are often remembered by the best-known name or by the employer. The first author of this one is Li, and the Why We Project cites it as Li and colleagues, with Huffman and Tokuda named as co-authors.

Key papers

Li, J., Huffman, S., & Tokuda, A. (2009). 'Good abandonment in mobile and PC internet search.' Proceedings of the 32nd International ACM SIGIR Conference on Research and Development in Information Retrieval, 43-50.

The paper defines good abandonment as a search query that the user abandons without clicking a result or refining the query, because the results page already answered their need. The authors analysed abandoned queries on PC and mobile search across three locales. They report that good abandonment makes up a large share of abandoned queries, that it is significantly more common on mobile than on PC, and that its main categories vary by locale and device.

They conclude that query abandonment should not be treated uniformly as a negative signal, and that search engines could encourage more good abandonment, especially on mobile, by improving search features and result snippets.

Only the abstract-level description was read for this page. Figures, methods and the list of categories are in the paper itself.

The paper is short, and its consequence is wide. It is the earliest source in the Why We Project's account of how click-based measures mislead, and it is cited here alongside the later studies of paid search at eBay and of long-term learning from ads at Google. Each of the three shows a different way in which a count and its meaning come apart. Li's paper is the one in which the count is the absence of an action, and that is the form most easily mistaken for failure.

Key ideas

The central idea is that the absence of an action does not have a fixed meaning. A person who does not click may have failed to find anything, or may have found everything at a glance. A system that scores both as failure will misjudge its best results.

The second idea is that the right reading depends on the device. The paper reports that good abandonment is more common on mobile. The Why We Project reports the difference without offering a cause, because the sources read do not give one.

The third is a design consequence. If answering on the results page is a legitimate way for a search to succeed, improving snippets and search features is a legitimate goal even though it lowers clicks. That is the authors' recommendation, not a tested result in the sources read here.

For readers interested in how products are measured, the idea generalises. Time on page, return visits and click counts each have a good and a bad reading. The page on Engagement Metrics sets out how the same count can mean different things, and Good Abandonment is the documented case from search. That generalisation is our reading and goes beyond what the paper claims.

---