|Formulating actionable theories in a post Panda universe|
If you're reading this you've likely read all the theories as to what Google may have done that has affected a full one eighth of all internet searches. Everything from duplicate product descriptions to changes in backlink value to other less tangible ideas have been thrown about and mulled over like a dog on a juicy bone. Except Panda wasn't so juicy for many, in fact many have seen their traffic dry up.
I've read all of the theories too and in as practical a manner as I could I've tested what could be tested with varying degrees of success. Frustrated, like many of you, I stepped back and tried to grasp the bigger picture and I've come up with yet another theory for you to mull over.
- 1/8th of internet searches were affected.
- Google dislikes spam in its index.
- Google likes interlinking its properties and features.
- Google has invested heavily in maps, places and various incarnations of local content.
Using just those facts my theory is that one thing in common to most spam sites is that they have a lot of product based content with a lack of a brick and mortar business address behind the site/business. Before I suggest that Google is giving more weight to sites with an identifiable business address I want to explain why, if that's true, it's not JUST a business address that matters. In true Google style it seems to be more complicated than that.
Blue widgets, lets assume my testing was about blue widgets. They are popular and many sites write about them and they are in fact a product that is part of most people's every day life. Sold everywhere, etc. Post Panda I wish you good luck in ranking a new page for the term blue widgets, it's near impossible. At first I thought that was because the competition was too extreme but I've found several product keywords that have relatively little competition, in the range of about 1.2 million results in Google.
The testing begins with these keywords.
Sure enough my blue widgets articles fail to break into the top 200 right out of the gate despite being in prime locations on established sites. Variations on blue widgets however, such as vintage blue widgets for example, are immediately in the top 20. Is that because there is less competition? No, it's not, there are 1.15M competing results for that term too. It seems that popular product names, when used in a manner to suggest a page IS about that product, have some sort of glass ceiling in place. The variation keyword does not share that glass ceiling despite being equally represented in search results.
So the theory is now that to rank for a product based keyword your overall site must pass a sniff test for that product and it looks like being tied to a brick and mortar address/store is one of the stronger authority signals. It makes sense that if you sell a product that you know a little something about it, which is something many content farm sites do not, especially with cheaply paid(or free), 3rd party outsourced content.
The theory: The level of authority required to unlock rankings on a product based keyword has been increased and a brick and mortar shop/business seems to be one of the no doubt many primary indicators.
I can't find any signals to disprove my theory and I can duplicate the effects it has predictably so while I'm sure it's not EXACTLY what Google changed it's tied to it somehow it is something I can act on. I do own a company that sells blue widgets and so I can let Google know I'm the blue widget guy for searches in the 123 anytown part of the world.
I think you are onto something here. As an online buyer, I know that I look for a business address and I rarely have confidence in any website that doesn't show me one. It's even worse if whois is private. The feeling of giving up credit card details to something made of only bits and bytes is not comfortable.
|Variations on blue widgets however, such as vintage blue widgets for example, are immediately in the top 20. Is that because there is less competition? No, it's not, there are 1.15M competing results for that term too. |
The number of results is not what it was in the old days. Adding a third word to a 2-word phrase can even increase that number estimate. So I can't get behind that as a dependable metric. Even on a result with supposedly millions of pages, I rarely can see the first 1,000 any more. Something abbout the Caffeine infrastructure shifted the way those numbers work - and they were not very intuitive even on Big Daddy.
I think the difference in "rank-ability" here is related to search volume - not to the number of potential results.
You have some good points back in November last year I was talking to a friend of mine from San Diego and he discussed how he was able to help a lot of his clients achieve great rankings for local results.
His idea was similar to yours but with certain additions, he was posting press releases on various sites and was including physical address of the business on each press release. The same for blog posts he was doing, he had the client name and the local business address and other credentials.
So far so good, let me tell you about 2 of my sites that got hit.
When my friend shared that topic with me I dug into some research papers on topic of Quality. (Earlier last week Bill shared a great post on his blog about quality as well).
I did my research and I made sure that I let Google know that my sites were maintained by real identity with phone number/address to reach and credentials plus other stuff.
But none of these had any impact my sites still got hit by the Panda update.
My take on all this is that somewhere G made a mess in some class or something. There hasn't been so much trash talk about Google Search or Google Adsense for so long. But if you check now everyone is complaining about how their business was impacted by recent panda update.
Even May Day didn't have so much of an impact that this last update did. As Aaron puts it on his blog Google is going towards paid inclusion system where you have to pay to get included in their "paid advertising" which will take pretty much everything from the search bar till 6th position.
From Bar till 6th position there will be Sponsor Ads and local ads and then places.
@tedster, I do have whois protection on all of my domains but that is to hide myself from SPAM EMAILS. Clients can reach me easily through my sites contact form either through phone, physical mail or an email.
|I do have whois protection on all of my domains but that is to hide myself from SPAM EMAILS. |
Yes, and that is one very real reason to use it. At the same time, when it comes to e-commerce, if I can't find solid contact information (something more than email) and then even their Whois is private, my "buyer's confidence" goes through the floor.
Are you suggesting that I take off the whois protection? Even though I have real physical address on our "about and contact" page?
Further none of my sites are e-commerce at least the ones that got hit by Panda update.
How about PO Boxes? I have a PO Box in Thailand that I list on the WHOIS. This is to protect me and my family from all the expat nutters that live in this country. My partners have offices in Hong Kong, Tokyo and Shanghai (real office addresses, not PO Boxes)... I list all four on the 'contact us' page.
That would satisfy me as a user - real, verifiable physical addresses are the kind of "comfort information" that I want to see.
Is there an accepted convention or format for showing an address and phone number, to make it more recognisable to a robot, such as an attribute/tag of some kind?
You could try hCard microformats. Google's rich snippet testing tool works with them, so they're probably a good bet. See http://microformats.org/wiki/hcard [microformats.org]
Tedster, I have a toll-free number on the ecommerce pages of my site. In the "About Us" page, I give an address, along with information about my corporation (when it was incorporated, etc). I don't hide anything.
Still, Google completely trashed the ecomm section of the site. There's certainly other factors at work, but I'm not sure how much weight to lend to phone numbers and addresses. Some of the sites that took my place are very anonymous.
There is no single factor present on every Pandalyzed domain. Rather it is quite a complex algorithm and we're sort of poking around at it. Not only that, but it is painfully clear that not every non-Pandalyzed site comes anywhere near being "quality" as a human would register it.
So this thread was asking for actionable theories. I think putting out every quality signal you can to visitors can only help - especially because user engagement metrics [webmasterworld.com] may very well be in play. After all, one of the Google questions was "would you be comfortable giving this website your credit card."
Sorry, Tedster. I wasn't putting down the contact information suggestion (although in re-reading my post, it sure sounds like I was).
Yeah, you did have me fooled there. That said, if just adding more contact information actually restored someone's traffic, my jaw would be on the floor!