Google Launching Payday Loan Algorithm 3.0 Targeting Spammy Queries

[Mentat]

---

System: The following 15 messages were cut out of thread at: http://www.webmasterworld.com/google/4676767.htm [webmasterworld.com] by aakk9999 - 5:06 pm on Jun 12, 2014 (gmt 0)

---

Google Launching Payday Loan Algorithm 3.0 Targeting Spammy Queries This Week

Didn’t Google just update the PayDay Loan algorithm a couple weeks ago? Yes, version 2.0 was launched a couple weeks ago.

Matt Cutts explained that this goes after different signals. The 2.0 version targeted spammy sites, whereas version 3.0 targets spammy queries.

Again, this is launching as soon as tomorrow and will look specifically spammy queries, such as terms like [payday loans], [casinos], [viagra] and other forms of highly spammy queries.

Matt did not clarify exactly what he means by spammy sites versus spammy queries, we will try to get more concrete details on this soon.

[searchengineland.com...]

[Martin Ice Web]

It is amusing to see Google pushes out one Algorithm Change after the other. It shows that their main algo isn´t able to match/compute the worlds Information stored on Websites.

And even spammy site should be catched by Panda? IMO a spammy page is a low content/shallow Content page?

[Mentat]

In my niche, since Spammy Algo 2.0 and Panda 4.0, I see a lot of "indexed search pages" as search results, full of spammy queries and a lot of pirated content.

I hope I will see a fix for this, because is getting ridiculous/unbearable.

[defenseup]

Why not Penguin then lol.. Come on already..

[aristotle]

Again, this is launching as soon as tomorrow and will look specifically spammy queries, such as terms like [payday loans], [casinos], [viagra] and other forms of highly spammy queries.

What is Google's definition of a "spammy query"? The examples [payday loans], [casinos], [viagra] don't seem to have much in common, except possibly they are disreputable or disgraceful or unseemly, in some peoples' minds. (although another possible thing they have in common is a lot of adwords competition.)

[EditorialGuy]

What is Google's definition of a "spammy query"?

A query that attracts spammers, maybe?

Not too spammers try to elbow each other out of the way for "mother teresa" or "history of borneo." "Casinos," "viagra," and "payday loans" are a different story, and it isn't too difficult to believe that Google might have a pretty good idea of what the most popular "spammy queries" are.

[Shepherd]

What is Google's definition of a "spammy query"?

I think if I were google I would be able to define a "spammy query" pretty easily using adwords data to determine high value keywords.

[Ecopac]

I don't get it, a query on it's own cannot be spam - surely only the websites that rank for that term can be classed as spam, so this update is going after spammy websites?

[EditorialGuy]

I don't get it, a query on it's own cannot be spam - surely only the websites that rank for that term can be classed as spam, so this update is going after spammy websites?

OK, so maybe Matt Cutts should have said "spam-prone queries." The point (IMHO) is that, if we're to believe what was announced, the newest version of the Payday Loan Algorithm will target search queries that attract spammers the way a picnic attracts ants. But--as always--the ranking decisions will be about pages and sites.

[aristotle]

Well if Google has a separate part of the algorithm just for "spammy queries", as apparently they do, then don't they need to be able to identify them in some algorithmic way?

[Shepherd]

don't they need to be able to identify them in some algorithmic way?

if ($adwordscpc>$x & $numofadvertisers>$y) {
$spammyquerry='yes'
}

[EditorialGuy]

Well if Google has a separate part of the algorithm just for "spammy queries", as apparently they do, then don't they need to be able to identify them in some algorithmic way?

Of course.

Something else to consider about the "Payday Loan Algorithm": If I were a gambling man, I'd bet that Google isn't nearly as worried about collateral damage when ranking results for "spammy queries" as it might be for more general queries. Aside from anything else, the Payday Loan Algorithm is probably a great test bed for new or more aggressive spam-detection tools.

[Shaddows]

If I were to go purely hypothetical...

Info queries get treated one way
ecom queries get treated another
certain uber-niches have special rules (e.g. health, travel, p0rn, news)

Now Spam gets its own micro-algo, it's own on-the-fly SERP-building rule set.

The final ranking is based on the indexed scoring of the page... how each of those scores are calculated would be the focus of V2.0.

Actually, I would think that is why Brands have been over-emphasised until recently. Whatever scoring criteria (e.g. domain longevity, traffic numbers) that represents brand-iness have been dialled up because they are unachievable for churn-and-burn sites. If there is a new separate algo for spammable queries, churn-and-burn sites will automatically become less of an issue for the rest of the world, allowing those "brandy" signals to be dialled back.

*****
As for identifying target queries... once you have a seed set of spammy sites, track what they rank for. Frequency analysis will give you the spammy queries. The advantage is, doing it that way means a query gets covered as soon as it gets a history of being targeted by spammers.

[aristotle]

I guess the real question is whether this impending refresh of the spammy-query part of their algorithm will actually get rid of any spam in the search results. Evidently they created this special part of the algorithm out of their frustration with their continuing inability get rid of spam with the main algorithm for all these years.

[Mentat]

Anybody detected any movement?
It's really strange that for my account, Google Analytics doesn't show any info/hour today.
Nada, 0 and there few hours left for this day.

[Ecopac]

I'm in the payday niche and nothing much happening at all. My traffic is up nearly 200% since the last refresh so I'm hoping nothing drastic is going to happen.

Been a quieter day today traffic wise but no movement in rankings. People must be looking after their money better today :-)

[SnowMan68]

I think Barry jumped the gun on saying today.

[rustybrick]

Matt said today, likely today.

[rustybrick]

Seems like It is rolling out now [twitter.com...] (as of 10 minutes of me posting this)

[Ecopac]

Keeping everything crossed I don't get shot down. Even though my site is whiter than white there's always a danger of getting caught up in this. #prayforecopac

[mirrornl]

yeh, looks like an old "dance"

[Ecopac]

Not seen any changes yet in the UK payday serps. The sites with the Uber spammy backlink profiles are still top..see what it's like in the morning

[BillyS]

I just hope that Matt's team understands that not everyone that writes about payday loans is evil. There are some of us that even warn of the downside of these loans.

No good deed goes unpunished....

[FranticFish]

Whatever scoring criteria (e.g. domain longevity, traffic numbers) that represents brand-iness have been dialled up because they are unachievable for churn-and-burn sites.

There clearly are other signals that can outweigh these though, perhaps dependent on query as you suggested, because the last time Google proudly announced they were cleaning up the payday loans SERPS they were handed their a** on a plate within a few days - see [seroundtable.com...]

Maybe this time they've decided to dial the 'freshness' aspect down?

[Ecopac]

Nothing has changed this morning from what I can see. However since the last update in the middle of May the serps have been cleaner than they've ever been. Haven't seen a spam site for a month now.

Just the heavily back linked sites ranking now, although they are genuine payday sites. No doubt penguin will deal with that.

So far so good really from my point of view.

[Shaddows]

There clearly are other signals that can outweigh these [brand signals] though

Sure, especially under the general algo. Imagine there are a bunch of signals. Some of these are usually really good indicators of quality, but are also gameable with enough effort. There are other blunt-instrument signals which are usually less helpful, but harder to game.

When people start spamming, you can dial up the un-gameable signals to minimise spam, but it impacts quality across the board. Or, you can aim for wide-based quality at the cost of highly spammed niches.

That seems to have been the choice for quite a long time, until the simultaneous (co-incidence?) Spam / Panda update. Suddenly, Brand signals had been dialled down on most niches, as if another tool had been deployed to clean the cesspit.

I would suggest Google actually got better at detecting spam. They then could dial down the Brand prophylactic. If I were Google, the next thing I would do is turn Brandiness back up on spammed niches, to further neutralise spamming attempts in those areas.

[Sgt_Kickaxe]

Sounds like a list of stopwords to me, keep them out of your page titles and you should be fine.

[Mentat]

No real changes and Analytics is blind :/

[SincerelySandy]

Whatever scoring criteria (e.g. domain longevity, traffic numbers) that represents brand-iness have been dialled up because they are unachievable for churn-and-burn sites.

I agree that they have been dialed up, wayyyy up. As for unachievable for churn and burn sites....
Months ago I got screwed by a well known company with a big brand. The company has been around for 10 years and has had countless mentions and links from big well established sites like amazon, forbes.....
One month ago I launched a new site on a new domain name. I post most of my focus into making the site look like the official website (without actually saying it or even coming close) of the company that screwed me.
I only have 3 links to the site and it has made it to the first page for most of it's targeted terms, including the name of the company that screwed me. It is listed above amazon and countless other much more established sites than mine.
The biggest thing I've learned about launching a new site these days is this... finesse. I waited about 2 weeks after launching the site before pointing a link at it. Then I watched it's positions in google and when it stopped moving up and down in the results, I added another link. Waited for it to stop moving up (and down) again, and pointed another link at it. After the 3rd link, it's on the front page.

[aristotle]

I agree with Shaddows. By creating a special separate part of the algorithm to deal with spammy queries, Google doesn't have to use the main core part of the algorithm for that purpose, and thus has more flexibility in finding ways to improve the general search results. Previously there was a conflict between efforts to get rid of spam and efforts to provide the most relevant and highest quality results. That's why they've never had much success at getting rid of spam in the past. But now they should be able to do a better job of it. And they should be able to improve the general results too. In fact it seems to me that this new approach could mean that Penguin and Panda might not be needed anymore.