Kill A Site via Akismet

[JasonD]

http://www.webproworld.com/webmaster-forum/threads/117356-So-Who-wants-the-Google-Scoop-of-the-year

[grnidone]

Seriously Jason.  What the fuck are you doing on the computer.  Take a break for a week so.  You've had a hell of a stressful few weeks/ months/ years, and you need a break.

[dogboy]

 

[littleman]

>Take a break for a week so.

Yeah, if that works for you.  I know we are all wired differently.  Some people need to keep at it when times are hard.

[Rupert]

Ok, so if its not Askimet:

http://searchengineland.com/google-no-we-don%E2%80%99t-use-akismet-to-catch-link-spam-117565

Then it is another source.  The step to Askimet was a jump too far.  But an interesting premise, that does not take away from the question, "how did Google know about the link"?

You always tell me Jason that negatives SEO is a real threat.  If this is right, now I guess it is starting to get mainstream.

[JamesR]

Links that appear quickly and then are removed after a short period of time (moderated) are a huge spam signal

[Gurtie]

does not take away from the question, "how did Google know about the link"?

quite so - on the basis of what was said, the link was never published or clicked on (blog owner mass marked batch as spam - infers he didn't start clicking through as he didn't even read the comments)

self hosted wordpress blogs - so no backend access.

so - afaik standard installs of wordpress don't publish the comment in the html of the originl post page UNTIL its published - the only person who see's it in the thread is the person who posts it, pending publication? I could be wrong on that - long time since I commented on any blog anywhere.

If that's right though, and Google 'saw' a copy of the comment in situ it can only have come from the person who made the comment - presumably via Chrome? If G have a copy of that and then have the published page html showing  that the comment wasn't the page, within a reasonable time, perhaps they could take that as a spam signal?

Or Google lie.

Or the original post is wrong.

[Rooftop]

Or the original post is wrong.

Seems to be the most common issue.

[Gurtie]

OR - he has a gmail account which recieves notification of comments?

Although that doesn't actually place the comment on the page so its a hell of a leap.

[Rupert]

OR - he has a gmail account

that would be a breach of privacy...  I guess Rooftops option is the best.  It is always easy to miss a small bit of the proof, and something like this could be viral.  Nice way of getting backlinks

If I were Google, I would have spam traps though.

[Adam C]

seen some speculation that Google's wording leaves some posibilities open.

just tracked it down... http://searchengineland.com/google-no-we-don%E2%80%99t-use-akismet-to-catch-link-spam-117565#comment-495155903

Google's reply, as others have stated, is also troubling. "we dont use Akismet to flag spam" could be read as:"we use Akismet to reinforce data on spam"

Think about it. Google has a hunch a website is over optimizing (Matt Cutt's words) and wants to verify that hunch. So they dive into the Akismet API and see how often the URL is blocked via Akismet as a check on their own blocking / spam data. and BAM! you have a great way of finding websites that are trying to optimize for keywords.

A bit tin foil hattish, but...

[Rooftop]

It would be far simpler to have lots of honeypot sites and just see who posts comments.  Don't feel very google though.

[Gurtie]

that would be a breach of privacy...  

hmm. would it? Automated reading of emails for spam detection any more than automated reading of emails for ad-serving?

to be honest it would surely be easy enough for them to see who (appears to) spam without setting up honeypot sites. 10 minutes online and you can find a hundred blogs with spammed out comments. Even Marissas    But none of these would explain the OP, IF they were correct in what they were saying

[JasonD]

> Automated reading of emails for spam detection

Google do this already. They define what emails are ham and spam for Gmail

Would it be a push to say they are covered in their ToS to scan any "new comment" emails that come into a Wordpress admin's email account for "Spam detection" of the on email kind?

[I, Brian]

I remember telling Adam a few years back that his anti-spam team should consider using Akismet for flagging purposes.

Then I figured that Matt Cutts probably had so much automated spam on his blog that he would already have thought of using some of of system using blog comments, whether through Akismet or not.

Either way, the link seems to suggest that loads of blog comments had been made, but the site was facing a penality because one or two blogs might be flagging the comments as spam. IMO it would have to be far more wide ranging as that - from my own experience, Google tend to allow for uncertainty, only looking to act when the signal is too strong to ignore. £10 says those "interractive comments" were using unnatural keyword anchors.

In the meantime, I've seen competitors that do some limited blog commenting *without* keyword anchors doing surprisingly well, though I figure it's part of a wider ranging link building campaign.

2c.