Google Search Is Poised to Surface ‘In-Depth’ Stories, but How Will It Pick?

Search results surfaced by "algorithmic signals."

  • Share
  • Read Later

Google's example of an "in-depth" result when searching on the term "censorship."

If you use Google News, you’ve probably noticed Google’s righthand sidebars, which, if you’re running the default (non-customized) version of the page, offer regional news based on IP address, a carousel of featured news sites with editors’ picks, further “spotlighted” stories and other multifarious context-sensitive info-capsules.

If you use Google Search and type something current, say “Edward Snowden,” instead of seeing his Wikipedia entry (the top or near-top result for nearly any search these days) followed by images and older stories, you’ll see a brief roundup of Google News-collated stories with times (published so many minutes/hours ago) beneath the heading “News for Edward Snowden.”

A new version of Google Search in the offing riffs on both of those features to bring you what Google’s calling “In-Depth” results, surfaced alongside the usual ones. Search on a given topic and Google will now designate a non-sidebar area directly in the search results comprising three sources, their names highlighted in green, with a sentence or two of teaser text to the right of each article thumbnail. The idea, says Google’s Pandu Nayak in a company blog post, is to service the roughly 10% (according to Google’s research) of people looking for “more than a quick answer.”

Google already applies an “in-depth” tag to stories that appear in Google News clusters, though what Google deems “in-depth” is often debatable, which raises the obvious question: Who — or what — is doing the picking?

According to Google, “In-Depth” draws its results from “algorithmic signals,” which is another way of saying “secret sauce” and leaving the specifics to speculation. You can influence Google’s algorithm by following the company’s guidelines: basically ensuring you’re following the latest indexing practices (as well, of course, as using Google services like Google+).

It’s not clear yet how comprehensive Google intends “In-Depth” to be, and that lack of specificity raises questions: Will it only apply to a handful of topics or eventually encompass every search term? If the latter, is it a veiled attempt to send this so-called 10% off to another corner of the web before they click the go-to Wikipedia entry? Are three articles sufficient? If the sources don’t cycle with each search, how long will they persist, and will the algorithm surface better material as it comes along, say a more insightful or informational piece that bubbles up days or weeks after an event?

Speaking of insight, how agnostic will the algorithm be when differentiating between vapid analysis by an established news outlet versus outstanding analysis by some no-name blogger? Is the algorithm weighted toward established venues? Is handpicking (or validating) involved at any step? Since we’re talking about a service self-evidently aimed at more inquisitive searchers, shouldn’t Google consider including Google Scholar results, too? And what about the other 90%? What if these “in-depth” results get in their way? Will people have the option to disable the feature?

In theory, something like “In-Depth” could become a way of preserving and surfacing important, well-conceived older material. But I’ve also seen Google News label plenty of questionable material “in-depth,” and even that term has problems: A voluminous screed by an ideologue or baggy extrapolative speculation about the latest wild Apple rumor might technically qualify as “in-depth,” meaning someone’s assembled a mountain of details, but assuming it’s useful information people are after, how does Google plan to differentiate between organized and factually reliable information and mere information (or does it)?

Just a few of the questions I have at the outset.