Irishwonder’s Black Hat SEO Blog A blog about blackhat, general SEO issues and other things related to the life on the web

Last week, everybody was discussing Google’s leaked AdWords score values. I guess enough guesses were voiced already about what those values might mean so I wouldn’t bother posting about it if not for one thing. I believe one more question that needs to be asked is WHY these values leaked.

Here are the two scenarios and it’s up to you to decide which one is more likely, if any:

1. Google is tweaking something - why else would their programmers touch this bit of code and leave these parameters slip into the open? As in, forgetting to comment it out or something.

2. Google is playing some PR game - as in, public relations, not PageRank. Google lets the scores leak to show something to all the SEOs out there, threaten them, keep them busy discussing it, or something.

Sure thing Google employees are human like all of us and Google is not immune to leaking these scores accidentally - but do you believe this was accidental or is there an agenda behind it?

… but I’ve been seeing discrepancies between different tools and toolbar PR on multiple sites - could it be an update?

You think I’m mad? Lately I have seen so many weird things in google.co.uk that are way too different from google.com that it would lead one to believe these are two separate search engines!

1. Why else would google.co.uk show ads for searches for something like http://www.somedomain.com/somepage.html? Especially since they are ads for the same somedomain.com? Sure UK googlers are greedier than their google.com counterparts and they are trying to squeeze every penny out of the poor advertiser!

2. Why else would a search on google.co.uk for a bit of page content in quotes (=exact match) result in two results: “Nothing found for your search in quotes but here is a result for your search but without quotes”?

Anybody else seen any more inconsistencies between these two different Googles?

Recently I have noticed that Google has become much more sensitive towards automated queries sent to it. It bans the IP very fast - much faster than it used to before. What’s funny is that the tool that got caught when I discovered it wasn’t even a site geenrator of any kind scraping the SERPs - it was a simple position checker Previously, it was possible to solve the problem by simply increasing the wait time between the queries sent to Google - not any more it seems!

Just another example of Google creating a problem for itself - if there was API for search still publicly available who would have queried the human interface SERPs directly?

As to the solution - I am now simply sending queries from multiple IPs instead of the same one

I’ve been looking through my server logs and this has caught my attention. The site in question just had AdSense placed onto it but no links have been pointed at it. (The site is not new, it existed for some time, then AdSense ads have been placed onto it)

What I saw in the logs is Mediapartners-Google crawling the site some time after the ads were placed (about an hour or so) - and then right after it, 14 minutes later, Googlebot coming over. Googlebot just asked for the home page - but that’s till interesting. Another thing worth noting is they both came from the same IP - 66.249.67.104.

Looks like this is all AdCents is good for nowadays, however. As a side note, I’ve had a quick look at referral products at AdSense (I DEFINITELY have no intention promoting these - I’d rather find an affiliate program in the niche that interests me) - and the funniest part of it was Google taking its geotargeting way too far. Namely, it assumes you will be targeting only the local users with your websites and you need a few extra clicks to change it. Dohh! Hello Google, ever heard of websites being just slightly different from traditional offline brick-and-mortar shops or whatever? I get visitors to my sites from places I have never even been to - moreover, sometimes I even build sites SPECIFICALLY to target the audience in such places! How much dumber can AdSense get?

Been searching for something that is not indexed by Google and came across something I don’t think I’ve seen before:

So now, instead of just offering to check if your search terms are entered correctly and additional options like go to the URL directly, Google fetches whatever it thinks is close… Not the most logical step is it?

The importance of each one of different factors influencing the SERPs in Google seems to vary across different markets nowadays - e.g. I constantly see SERPs where PR6 pages go way lower than PR0 or PR 2 pages (old news of course, PR not being a major factor any more, at least in the form available to the general public), or I have seen SERPs where 100 links beat 10,000 links, and I have even seen SERPs (gasp! what would Jim Boykin say? ) where 2-year-old sites were beating 10-year-old sites to the top spots.

But here’s one example of SERPs where age factor is definitely overplaying to the point where it hurts the quality of these SERPs, as described in this post by Affiliate Money Maker:

I headed to Google and ran a quick search. Google returned almost 9 million results for “send a fax by email” - wow that’s impressive. What have we got here?

The first one was TPC Fax (tpc.int) - a site last updated in 2001. You see, in the old days of non-monetized Internet there was a community initiative to support sending faxes by email through privately operated printer servers. This was a free community service supported exclusively by good will of people participating in this project. Well that was a long time ago. I don’t know how reliable or widespread it was, but currently it is not working any more.

Digging through a bunch of FAQ’s and equally old sites that are of no use any longer,<...>

It is pretty clear that the search was meant to be for sites actually providing these services - and what comes up instead? A bunch of aged but practically useless results… Not a good job, Google.

Ever wondered what happens once you send that ping over, be it for a blog or otherwise?
Here’s what:

*

/2007
Http Code: 404 Date: Feb 07 22:57:56 Http Version: HTTP/1.1 Size in Bytes: -
Referer: -
Agent: Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)

Yea you guessed it, this wasn’t a blog. It certainly gives me a couple ideas…

No it’s not your general advice on using subdomains vs sunfolders. Notably, Matt manages not to say anything meaningful and up to the point till half into his post - and the real juice comes in comments:

Google will be less likely to show (say) 7-8 subdomains from one domain for some queries

- whatever that means. No Blogger blogs by thousand in SERPs? - I doubt. Spammer detection, life threat to spammers using subdomains? - hmm I still suspect it would take a manual review for that…

I do like a bit of conspiracy theories every now and then. Was doing a search in Google today for something related to .edu domains - and my attention got drawn to a bunch of Google ads that came up with the results opf my search. All those ads read along the lines of “get real .edu links to boost your site’s rankings” and such. That got me thinking: sure it’s not the first time Google is caught talking negatively of things and still displaying ads for them, but could this be part of their plan in this case? I mean, there are several things they can potentially be doing with these ads (besides ripping thie immediate financial benefits):

1. Obviously this is an easy way to locate link selling networks - and if I was running one I definitely wouldn’t have been advertising it in AdWords, no matter how much I want immadiate traffic an sales.

2. OK this one is a bit too far fetched - but what if…. We all know how much info Google is already collecting on its users - imagine that a person clicking an ad is signed into their Google account… First Google tracks if a click resulted in a conversion (i.e. signup) on the link selling site. Then, Google checks the user’s history to see what domain(s) interest him most, thus assuming some of them at least may belong to him. Next, check within some period of time (say a month) if any of these domains acquired any new links since the conversion on the link selling site. All of this happens in bulk of course. Next, add two and two - and you got a complete picture of the network where the link dealer is selling links. No need for footprints even. AdWords as a honeypot.

I am paranoid and it’s a fact but it doesn’t mean they’re not after you.

Conclusions:
- I know - and always did - to make sure I am not logged into any Google acct when searchign for anything important
- I know if I was buying links for my sites it’s best not to buy from networks advertised in AdWords - or at least avoid clicking the links to them directly from the ads
- I know what I would be doing if I was running a link selling network
- I know it’s best not to post it here or elsewhere publicly in case Google folks are still reading my blog
- I will now know if Matt Cutts is still reading my blog hehehe - hey Matt