61

Today, I was searching on Google for a resource on viewing the JavaScript console in various browsers. As it turns out, there was a good overview on Webmasters.SE:

Google search with John Conde as the Google-listed author

"Thanks, John Conde!" I thought to myself, as I clicked through to the excellent comprehensive resource.

But wait -- John Conde didn't ask this question, nor did he supply any answer. How, then, did Google come to think that Mr. Conde was the author of any part of this post?

Then at the bottom, I saw:

"protected by John Conde" banner

One of the mechanisms to establish Google authorship in a search is:

  1. Make sure a byline containing your name appears on each page of your content (for example, "By Steven Levy").
  2. Make sure your byline name matches the name on your Google+ profile.
  3. Verify you have an email address (such as stevenlevy@wired.com) on the same domain as your content.

It appears that the "by John Conde" in "protected by John Conde" is read as an authorship byline by Google. (This may also be true of the first close-voter in a closed question's "closed as off-topic by John Conde,..." but I haven't observed this yet.)

Thus, it seems that to reproduce this behavior:

  • a question must be protected by a user who has a Google+ account
  • that user must has a username that matches his real name on his Google+ profile
  • the user must have a @stackexchange.com email address registered with Google+ for ascribing authorship (probably not necessary; see edit)
  • the user must have reciprocal rel="me" links between his Stack Exchange profile and Google+ profile

I naturally don't think this is a very serious or high-priority bug, but it might possibly be solvable by using some sort of <meta> tags or link rel attributes (though I haven't looked into it). The worst harm this bug is likely to do is misrepresent the list a person's authored posts when someone views the person through Google search as an author.

EDIT

@random notes in the comments that John Conde probably does not have a stackexchange.com email address. This seems strange, because Google lists only two methods for showing authorship in search results. The first method requires a confirmed domain-specific email address. The second requires reciprocal links between the authored post and the user's Google+ page, but neither the post nor John's Stack Exchange profile link to his Google+ account.

I'd guess that there's some slightly undocumented method being used here to determine authorship. John probably has reciprocal links between his Google+ page and his personal site, http://johnconde.net, which is reciprocally linked to his Stack Exchange profile. The link path goes:

Protection banner -> SE profile <--> Personal site <--> Google+ profile

The link in the "website" field of a user's SE profile uses rel="me", and John's personal site includes a reciprocal rel="me" link back to his Webmasters.SE profile. According to Google's writeup on the use of rel in describing authorship:

The reciprocal rel="me" links tell Google that the profiles... represent the same person.

My best guess is that:

  • because Google equates the SE profile and the personal site, and
  • because Google equates the personal site and the Google+ profile,
  • therefore, the mistaken-byline link to the SE profile implies authorship by the owner of the Google+ profile

This might work just as well without the intermediary personal-website step, if the user has reciprocal links directly between their SE profile and Google+ page:

Protection banner -> SE profile <--> Google+ profile

Ultimately, it seems this comes down to Google misreading the "protected by [name]" as an authorship byline, which is obviously Google's fault, but might be within Stack Exchange's power to fix, with the proper application of SEO magic..

  • 2
    I read that as "John Code" – Cole Johnson Aug 13 '13 at 1:59
  • 5
    Pretty sure Conde does not have a stackexchange.com email address nor would it really matter here since the first two points satisfy – random Aug 13 '13 at 2:04
  • Sounds more like a Feature Request to me, no? Not actually a bug with SO/SE... – Joe Aug 13 '13 at 13:48
  • 2
    @Joe Well, the current behavior is a bug, but the question is on whose side and whether it's easier for SE to fix that or just have Google tweak their algorithm. – slhck Aug 13 '13 at 13:50
  • Its a bug somewhere, possibly not an SO side bug though – Richard Tingle Aug 13 '13 at 13:51
  • Right, but 'bug' here is for reporting bugs with SO/SE - this isn't something SO/SE is doing particularly wrong, it's something google does that has an unintended effect. Hence 'feature request' since you're really asking for SO/SE folks to make a feature change (ie, SEO change). – Joe Aug 13 '13 at 13:57
  • 1
    @Joe I certainly see what you're saying, and I agree it's very much a borderline case. I'd invite a mod (or a developer who might one day work on fixing this behavior) to re-tag as necessary. – apsillers Aug 13 '13 at 14:01
  • I reported this on Webmasters Meta, too: Google displays wrong author info in search results – unor Oct 2 '13 at 12:57
  • Everything is already here on Meta. I just discovered this for a different search term and was about to post this as a "bug", only to realize it had already been posted. I +1'd this, but I'm not sure what kind of answer are you looking for? – Amal Murali Jan 28 '14 at 13:34
  • Would rel="nofollow" on the profile link in the protect and close notice boxes help Google distinguish a post's author from its moderator? – Damian Yerrick Nov 8 '14 at 16:37
8

We have just deployed a set of schema.org QAPage compatibility updates across the network to better inform Google (and anyone crawling us really) of what's what, who's who, etc. This includes clearly indicating the author's name.

This should be a behavior that doesn't repeat with their current crawler setup and the information we're being explicit about.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .