Panda Update On Ad Ratio and Placement

[hairresources]

One of the factors mentioned in the info about the Panda update is having too many ads on a webpage in relation to the content on a page. I have noticed in several of my sites that weren't hit as hard did not have any ads at the top of the page (when viewing the actual code for page)- at least not before several paragraphs of content.

I have been removing all but one ad unit on each page of several sites affected by Panda and placing this unit further down the page- still ATF but below some content.

I am seeing some competitors ranking on page 1 that have 3, sometimes 5 ad units per page with 2-3 of these ads appearing ATF I am wondering if moving ad placement and removing ads is the way to go?

[Planet13]

What do the INBOUND link profiles of your competitor's sites look like compared to yours?

[hairresources]

Inbound link profiles of competitors are very similar to mine. Some have even less inbound than my site.

[Planet13]

Well, there is really only one (or two, depending on what day it is) webmasters who say they have recovered from Panda. One of the threads is here:

[webmasterworld.com...]

I would read through that thread and see if anything there rings a bell.

Since I don't do adsense, I can't unfortunately give you any tips myself on experiments with ad placement. I am sorry about that.

[tedster]

Ad placement is apparently only one of many ingredients in the Panda recipe, so I doubt that there's only one "way to go". It's more likely that there's a combination of factors, and the core of those factors would be the quality of the content itself.

Overwhelming ads can sometimes be a secondary sign of a website that was essentially created only to carry ads, rather than to serve content that has solid value. In my opinion, that is why Google's algorithm picked up on ad ratio and placement in such a big way.

[dataguy]

My AdSense rep swears that ads above the fold aren't really an issue. Then again, he was advising me to place more ads above the fold when Panda 1 struck.

[potentialgeek]

I guess Suite101 has moved their aggressive before-the-fold ads after being Pandalized. Not sure if it helped.

Ad Ratio is probably a late signal judged in comparison to early signals such as text quality, article length, uniqueness, site authority, etc.

In the list of 23 questions submitted by Google's Blog after Panda, ad ratio only came up once, and that was towards the end.

When Google asks the question, "Does this article have an excessive amount of ads that distract from or interfere with the main content?" it is probably talking about ads which appear early and make it difficult to find/read the content without scrolling.

Some Adsense ads on Made-For-Adsense sites fill up most of the above-the-fold space (excessive) and/or are stuck below the page title and above the first paragraph (distracting), or simply butt into the early text.

Google has the technology to preview landing pages and any other page to see if ads are excessive/distracting. But it probably requires a lot of resources to capture and analyze the images.

[tedster]

Ad Ratio is probably a late signal...

Given the way that the Panda algorithm was constructed, I think the fact that it emerged as a signal at all is pretty telling. It seems to say that Google leans rather heavily on their virtual page simulation technology -- and that data that was already collected in this vein was a good part of the data that the learning algorithm ranged over in building the new algo.

Since the whole idea of machine learning is to create a predictive algorithm, I've been wondering why an 86% success rate in prediction was considered good enough to go live. The only possibility I've come up with so far is that it takes a full blown release to allow further fine-tuning and machine learning.

And since the main target was shallow content, it seems a bit surprising that a layout factor would even have emerged at all. I suppose that content farms would naturally be manipulating the layout to maximize ad exposure and income.

But then, the complexity of the Panda algorithm shows itself - it's clear from many counter-examples that ad ratio and placement alone do not trigger a devaluation.

[freejung]

since the main target was shallow content, it seems a bit surprising that a layout factor would even have emerged at all

For this sort of predictive machine-learning algo you don't need causation, right? You just need correlation. Even if ad placement has very little to do with whether people actually like the site on a case-by-case basis, if the type of sites they are targeting tend to have more aggressive ad placement than others for whatever reason, that will still emerge as a factor.

I bet over time we'll discover lots of factors that matter to Panda that no human would ever consider to be directly related to "quality," yet that's what emerges out of the math.