Matt Cutts Announces "Above The Fold" Algorithm Launch

[tedster]

In our ongoing effort to help you find more high-quality websites in search results, today [19 January] we're launching an algorithmic change that looks at the layout of a web page and the amount of content you see on the page once you click on a result...

This algorithmic change noticeably affects less than 1% of searches globally. That means that in less than one in 100 searches, a typical user might notice a reordering of results on the search page.

- Matt Cutts

[insidesearch.blogspot.com...]

[onebuyone]

tedster, I think there is other reason for this. Panda is not real time (yet), because that's how Google wants to defend against reverse engineering.
Is it computotionally intensive? It shouldn't be.

[tedster]

I have pretty good reasons to think you're wrong. Panda is not just a complex data scoring of a massive number of web pages - that's just the "data refresh" area of Panda. The core algorithm itself is constantly evolving through machine learning on a massive scale. This means that what specific metrics it uses to create the scores (and the formula for how those metrics are combined) are not the same from iteration to iteration.

There was just one Panda data refresh that did not involve a new version of the core algorithm and that was during the week of Jan 19-24, 2012. Members (including me) here had trouble pinning down the date any better than that, since it faded into all the other ranking changes that month.

Not only is this a vast machine learning process (done on a scale that even Google never before attempted) the learning process is being done over a moving target: all the web pages that are continually changing. Even IBM's Watson doesn't have this kind of challenge.

If they just want to protect the Panda algorithm against reverse engineering, all they would need to do is not announce which ranking updates were from the Panda portion of the total algorithm.

[guggi2000]

@tedster @onebuyone
I agree with tedster, I am sure it is computer heavy. But let us stick to Page Layout Algorithm: I think one can easily understand that painting the layout of each page (!) needs more CPU than just reading and parsing it as plain text. Google has already suggested that it is a manual update.

But, all this DOES NOT MATTER. The REAL QUESTION that has to be asked is when they will/did rerun it. Because if it has been rerun and people did not see changes, then they were 1.probably hit by something else from the beginning (i.e Panda in mid-January) or 2.their changes were not sufficient (still ad-heavy) or 3. their pages have not been recrawled yet.

Since I have not seen anybody recover from it yet I still have hope that I was hit by the Page Layout algorithm and not by Panda (which would be 10 times worse).

[Pete123]

I was hit Jan 19 with about a 50% drop on my main website and like most folks, removed all the above-the-fold ads but have seen no improvement in traffic, despite everything being recrawled weeks ago. Never been hit by Panda before, actually been helped. Since Jan 19, traffic has crept up maybe 10%, typical of what happens when a Panda is run, but of course it's still barely over half of normal.

I also have a different very small low-traffic website which had Adsense ads last year, which also crept up about 10% with each Panda. Last year, I decided to remove the Adsense ads for fear Google would think it was a MFA site and it would endanger my main site's Adsense account. Within a couple days, traffic dropped 50% on the little website, for no discernable reason. It stayed that way for about 80 days, then suddenly spiked up to normal again, and has stayed normal ever since. I added Chitika ads after it got back to normal, and no change.

Now it seems beyond strange that I got an 80-90 day penalty for taking off Adsense ads on the little site, but that was the only change onsite or off, and none of the Panda runnings last year had that kind of dramatic effect, up or down. It really seemed like a timed penalty. The little site wasn't affected on Jan. 19 because it had so few ads.

So I'm keeping my fingers crossed that the above-the-fold slap on my main website contains a timed penalty also, which explains why no one has seen recovery yet, and with luck, something exciting will happen around the middle of April, or whenever the timer ends.

[smithaa02]

The homepage of our our sites dived about 7-8 spots in mid-march. The homepage was ad-heavy (the sub-pages were not and weren't affected). However, since this happened mid-march, is it safe to assume that this would be too late to be an above-fold content penalty?

Also...how on earth does google know the difference between an ad and say glorified navigation button or picture? Are they parsing known link profiles? There are flash sites we keep an eye on as well as wordpress sites with mega slideshows at the top...they weren't hit.

I'm almost thinking this algorithm thing is a ruse..and google is using EWOQ or a similar human review team to determine whether a site it too ad heavy or not. This would also explain why many sites haven't bounced back yet...the human reviewers haven't got to them.

Determining how ad heavy a site is is EXTREMELY difficult with googlebot because you're messing with the javascript, foldable content, css layers, flash, and legit graphical content. Only humans IMO (at this point) would be able to tell the difference unless google cheats and goes after say sites with fingerprints with third-party ad services (like adsense).

Or perhaps it is a combination of the two... Googlebot flags a suspected ad heavy site for human review and the humans actually assign the penalty.

[FrostyMug]

I haven't seen any posts about recovery, but maybe it was highlighted in another topic?

I'm still in penalty, since Jan 19th, still only 50% of traffic.

Since the changes I've made in first week after penalty [reduced number of ads, aligned site to left, etc], I have not touched it at all.

Any update?

[BillyS]

FrostyMug -

We were hit with this penalty too. The interesting thing is our content is very high on the page. Our information starts around 110 pixels from the top. We had a search box above the content, which we moved to the side but we've seen no recovery either. It's just remarkable. Our navigation is on right, which is unusual and I'm wondering if that's throwing off Google.

I'm thinking these penalties do not go away easily. It's like getting thrown in the sandbox again, it could take months to recover after a fix is in place.

[guggi2000]

@FrostMug No recovery here either... "a few weeks" became months. I cannot rule out that another penalty was imposed on the same day.
However, it seems no one has reported recovery so far, which keeps me in the wondering.
By the way, looking at Google analytics the "crash" started on Jan 19, at exactly 12:00 West Coast time.

[freeport]

Just listen to yourself.. you are jumping ropes for google here.
Changing layout of your own property just to please the big G ?
Do you know how pathetic this sounds?
Lets start sending flowers to G , and don't forget to trow in box of chocolate.. we would not want to throw off Google,would we?.


"We had a search box above the content, which we moved to the side but we've seen no recovery either. It's just remarkable. Our navigation is on right, which is unusual and I'm wondering if that's throwing off Google. "

[BillyS]

Do you know how pathetic this sounds?

Quite frankly, the only thing this penalty made us do was take another look at our layout. I actually like the new location of the search box. So a big thanks to Google for making us think. :)

[deadsea]

I discovered something interesting in google webmaster tools. Under "Your Site on the Web" -> "Search Queries" I used to rank #1 for some queries that get 100,000+ impressions per month. Now if I sort by "Avg Position". I see that I still rank #1 for lots and lots of terms, but the top terms that I rank #1 for have at most 3000 impressions per month. It makes me wonder if this is because those terms are less competitive or if Google applies to "too many ads" filter to the SERPs only when there more more than 3000 impressions a month.

[smithaa02]

Deadsea...interesting...I might be seeing the same thing.

We got clobbered for our money phrase which was 100k+ a month (we recovered a bit then dived when penguin rolled out). Not sure if it an above-fold thing or a penguin thing...

At any rate...most of our results didn't get touched...only our money phrase to our home page. It's as if google has a completely different algorithm for 100k+ searches.

[guggi2000]

Any recovery for anyone?
We are still down 40% since Jan 19, 12:00 PCT...

[gouri]

Is there anyone who was not affected by the above the fold algorithm when it rolled out in January but feels that they may have been impacted by it later on?

I know that this may be hard to determine but maybe there are some who might be able to.

Has the percentage of space above the fold occupied by ads that causes a site to be affected by the algorithm changed?

Has this algorithm been made a part of Panda or Penguin and is now one of the factors in either or both of these algorithms?

Has anyone not impacted by the above the fold algorithm in January who had ads above the fold at the time moved them and seen any changes as a result?

I would appreciate your responses.

[topstar]

Is this algo being run regularly? I changed the design of my website and took a hit the next day. I might never know if it was the AtH or just a layout change in general that caused it.

[tedster]

As far as I can tell, the ATF algo is folded into the near-real time total algorithm. Only Panda and Penguin depend on a periodic re-run. It is very strange, however, that no one is reporting any recovery.

[realmaverick]

I was hit on Jan 19th, UK. I've never been hit by Panda in the past. It's rather unfortunate that both ATF and Panda update were released alongside one another. As it makes it much more difficult to know 100% which update we were hit by.

I have a max of 1 ad per page, so I can only assume I wasn't hit by ATF, but I've never been 100% sure.

[Kendo]

Re the "affects 1%"... does this mean that when we find tat we have been screwed that we are supposed to assume that because it only affects 1% that we are just unlucky?

Otherwise it's not the most idiotic suggestion that Google has ever made, but it's on par... to being so simplistically naive that it's incredibly funny.

[deadsea]

I'm firmly convinced that Google is not able to measure content effectively for my site. My content on the pages that was hit is a block powered by javascript. Some folks have indicated that image or flash content above the fold is ok. I'm seriously considering converting my site from javascript to flash to try do demonstrate to Googlebot that there is content there.

[realmaverick]

I've just looked at my analytics on 18th, 19th and 20th Jan. My traffic remained solid until jan 19th around 10pm GMT. But only ended slightly down. 20th ended 25% down, with a clear drop on an hourly basis.

Despite several people trying to convince me that it couldn't possibly have been the ATF, looking at my traffic graph, it has to be the ATF. But with no ads on most pages and 1 ad on others, it can't only be ads, this update is targeting.

Especially as sites in my niche with many ads above the fold, aren't effected.

Google is a mysterious beast.

[realmaverick]

@deadsea do you happen to have an exact time that you notice the drop?

[deadsea]

The first hour that looks down is Jan 19th at 1900 GMT. It drops hourly over the next 4 hours. It is down to its final level by 2300 GMT.

Here is the graph (in EST): [i.imgur.com...]

Edit: I realize I labeled 1900 GMT as 1800 GMT on the graph.

[realmaverick]

Thanks deadsea, very similar to what I experienced.

On the surface it appears a very straight forward update and one that can be easily fixed. However in reality that isn't the case.

I too haven't recovered, despite only having one ad on the page. I'm wondering if those sites that cannot seem to recover, actually share anything in common. My suspicion is that sites with low text content above the fold AND an ad (mine is 300x250) are struggling.

It would be useful to figure out, exactly how they're calculating the penalty. (wouldn't it always).

[guggi2000]

Same for us: Jan 19th, 2100 CET (12pm in Mountain View, California).
No recovery despite complete redesign.

[realmaverick]

I've read a few articles by webmasters of webmaster type blogs, claim to have been hit and then had sudden recoveries after making the suggested changes.

But I'm wondering whether their entire post was fabricated, so that their 'advice' has more punch.

Has anybody here at WebmasterWorld recovered? Or do you know anybody who has?

[Pete123]

No sudden recovery here. My traffic has gradually crept up to about 3/4 of what it was before Jan. 19, from an original drop to 1/2. But the increases have been gradual, with small ones that correlate to the Penguin/Panda updates, similar to what occurred last year whenever Panda was run. My other site that wasn't hit Jan. 19 has increased about the same, so I attribute the increase to Penguin/Panda, rather than any kind of recovery specific to the above-the-fold problem.

I removed all above-the-fold ads within a week or so after Jan. 19 and replaced them with one below-the-fold ad only.

[Thaparian]

Has anyone tried removing ads all together? Recovery after ad removal would confirm if the drop was related to ads, if no recovery then something else is wrong.

[guggi2000]

We recovered on October 9, 12:00pm. Traffic went up by 40%... (to previous levels).

In short:
January 19: Google announces "Top Heavy" (Above the fold) update... our site goes down by 40% at once (exactly 12:00)

October 9: Google announces "Top Heavy" update... our site goes up by 40%, exactly at 12:00

FYI: We removed the ads at the end of January, and constantly moved them down a little (by a few pixels), each month.

My personal conclusions:
- It is a site wide penalty
- Google ran the update just once after 9 months
- I believe that all the changes that I have made to the site during the past 9 months did not have any influence at all. Removing the ads and waiting probably would have done the job. However, I cannot be too sure, because I added content as well, sightly played with ad position and removed ads from pages that had thin content.

Can you share your experience too, so that we can compare and optimize?

[Shaddows]

@guggi2000

Just seeking clarity.
Did it drop 40% then return to full traffic? Or did it drop 40% then increase 40%?

If the latter, you would now have 84% of your original traffic.

For you to completely regain former traffic, you would have had an increase of ~67%

[guggi2000]

@Shaddows it returned to full traffic, 40% down and 40% up of the initial traffic.

I thought it would be too complicated to get into the math.
Thanks for pointing it out.

[TheMadScientist]

I thought it would be too complicated to get into the math.

Please, always get into the math ... We need way more of that around here again. It's been dropping for years and is definitely missed by some of us.

[deadsea]

I lost 40% of my traffic on January 19th, and gained back about half of what had been lost on October 9th. In my case, the ad layout was not very different at all. I still have some ads above the fold (one or two depending on the page). My ads are in prominent positions (often to the left of the content where users see them first). The main (interactive javascript) content is above the fold in a brightly colored area to make it obvious to users.

[Dlocks]

I didn't see any increase at or after October 9th on two of my websites that were effectet on January 19.

Although I removed 33% of the ads and placed the rest below the fold in February (I'm only using AdSense thus since February only two AdSense block below the fold).

In my case the two affected websites are Flash game websites. On the play game pages there is a short description and the Flash game itself above the fold. Below the fold I have other games (thumbnails and short descriptions).

Makes me wonder if Google conludes that a large flash movie of 728x530 above the fold might is not userfriendly (as there is almost no txt) and perhaps 'thinks' that it could be an ad.

For Flash game websites this wouldn't make any sense. You The visitor doens't want to see the actual game on a "play ... game" page below the fold. Because then the visitors needs to scroll down to see the content (the game) he is looking for. And the whole above / below the fold thingy is all about visitors not having to scroll down to see the content he is looking for. :P

But then again, at this moment I don't see any other option.

[Gimp]

Adsense sent me a message telling me that 67% of my ads were below the fold. It went on to tell me how much more money I might make by moving them above the fold.

And then it said moving them up would not detract from a good user experience and referred me to a video that showed many different recommended ad placements.

[Whitey]

UPDATE: Google: Our AdWords Ads Wouldn't Trigger Page Layout Penalty

In January 2012, Google announced the page layout penalty and then updated it in October 2012, it basically sets to penalize sites with too many ads, too much in the way of your organic content.

As I covered at Search Engine Land, the question was posed to Google's Matt Cutts about Google penalizing their search results page because there are so many AdWord ads pushing down the organic search results.

Matt responded two ways, the second more interesting:

(1) Google doesn't index their own search results.

(2) If they did, the ad to organic results ratio and layout would not trigger a page layout penalty.
[seroundtable.com...]
Full SMX interview with Matt Cutts and Danny Sullivan here : [searchengineland.com...]

[Whitey]

Matt Cutts just tweeted a comment onto the above article :

an important point missing from the write-up is looking at all the pages on the site, not just single pages.

[ColourOfSpring]

If they did, the ad to organic results ratio and layout would not trigger a page layout penalty.

So....it's fine to have 80% of your visitors see 90% of above the fold content as ads? OK, Matt.

[deadsea]

His statement makes me think that the ratio they are looking at is:

Amount of ad space above the fold / Total content space on page

Instead of what I had assumed earlier:

Amount of ad space above the fold / Amount of content space above the fold

[ColourOfSpring]

His statement makes me think that the ratio they are looking at is:

Amount of ad space above the fold / Total content space on page

Instead of what I had assumed earlier:

Amount of ad space above the fold / Amount of content space above the fold

If a page is predominantly made up of ads above the fold (for most screen resolutions), then it strongly suggests the purpose of that page is to sell ads. If the purpose of the page is to NOT sell ads, then it's a poorly laid out page.

[hitchhiker]

So, according to what Matt said:

Take the height of your ads / divide by the height of your page when fully rendered. Match the SERPs ratio, and you should be fine.

The ratio (for a 3 block, heavy popular phrase ad block) is 0.2

The ads on a google SERP take about 20% of the overall page height as measured by Chrome.

If that's true...

Unfortunately I decided that reading posts was a little easier with a wider 'template'. So all my content is now 'shorter', therefore my ratio is higher.

[ColourOfSpring]

Take the height of your ads / divide by the height of your page when fully rendered. Match the SERPs ratio, and you should be fine.

The ratio (for a 3 block, heavy popular phrase ad block) is 0.2

The ads on a google SERP take about 20% of the overall page height as measured by Chrome.

If that is what Matt Cutts is suggesting, he's breaking one of the most fundamental rules of what creates a good user experience i.e. summarise the content of the page above the fold. Just seeing a block of ads above the fold is a very bad user experience.

And if it's true that the actual length of the page is important, expect to see even more filler-rubbish populated on pages just to get ratios down....

[chalkywhite]

So my nice powerfull dedicated server that cost $$$ and makes the site uber quick makes no difference to my ranking. AAAGh. I thought site speed was important? seems you just need to make sure its not absolute dog only.