A few years back, we in the seo community noticed the Google SERPs taking on a new look. For most searches, exact match domains(EMD's), seemed to dominate the first page. For the next few years the affiliate markets boomed as long tail EMD's could garner the lion's share of the search traffic within a matter of days with minimal effort.

Blogs and forums lit up like Grandpa's birthday cake with advice and tips on getting to the top with EMDs. They are still a significant part of the placement process FOR SOME TERMS but there can be no question that in 2010 the astute SEO began noticing a change.

One incident kind of uncharacteristically tipped Google algorithm hand. BMW. 

They had gotten their cyber hinny paddled for some of the most elementary and easy to spot SEO techniques known to the industry. They pretty much ran the spam gamut from A to B and it was almost like they were daring Google to do something about it. So Google did.

However, BMW cried foul to the print digerati and within a matter of a fortnight BMW was restored. A kind of un-handjob,

Because of the publicity propagated by BMW themselves, it didn't take long for SEO's all over the planet to become aware of the event. Well what made the changes that some of the more astute practitioners had seen for months suddenly become common knowledge. That common knowledge was, yes, Google's algorithm was heavily favoring large brands for a LOT of keywords.

Now brands and the preferential treatment they got, both in terms of favoritism in SERPS and in the leeway given for what other sites would get whacked for, became a hot topic.Now some 20 months later that trend has only gained steam and for just about any consumer goods  search, we can all find the first page filled with the Wal marts and JC Penneys of the world often with obviously less relevancy than many EMDs and just better, more relevant, authority sites.

How to brand your affiliate site became the fodder that fed many a blog post. BUT the general consensus was, and remains, that a small site simply can not compete with a big BRAND site for competitive consumer goods and services searches.  As an industry we've all pretty much conceded defeat to the money, reach and influence of big brand sites. BUT what has always bothered me is that I know a computer spider doesn't know a brand from a boiled biscuit,. AND an algorithm could only determine a brand based on what the programmer wrote into it as filters and/or signals. So my question had always been, what defines a brand to a search engine?

Lets put ourselves in a programmers shoes for a moment. It occurs to me the FIRST question a programmer might have is :

Why would I care what a brand was?

In some boardroom, (probably with a snapple vending machine and a foosball table in it), there was a discussion as to why and how to identify and add significant relevancy scores to brand sites. This would have happened long before any coder actually got involved.

It helps if we remember that contrary to popular misconception Google is NOT a relevancy delivery machine as much as it is an ad delivery machine. Of course it wants some degree of relevancy or rather the perception of relevancy to keep the eyeballs of their loyal fan base returning and downloading more and more helpful tools. They have to have someone to show the ads to!  BUT free searchers do not put Maserati's in the parking lot. Paying advertisers do!

A relevant ad is one that generates a higher click through. So, why brands?

The glaringly obvious answer would be they pay more BUT that is NOT it. They aren't paying for organic results. They are paying per clicked on ad! NOW we see the only real motivation to provide relevancy. The more clicks the more revenue generated. BUT that still doesn't answer why big brands provide more relevancy.

ANSWER: big brands provide more data!

They know the consumer is more likely to trust a site and follow thru right up to the UPS guy ringing your doorbell from established branded sites than from thin content, triple hyphenated domains. By leaving the most relevant sites based on something as simplistic as page rank at the top we get far too many abandoned shopping carts, back buttons to the search page, searchers getting  re-directed or too many people hitting adsense ads giving too many false positives. BUT send  them to Macy's for women's fashions and they will get a much greater subset of data from actual purchases and from people viewing you tube to see the product in action, to gmailing their friends sale notices to going to facebook and complaining about the service. THAT"S WHY THEY WANT brands. To gather more data from more people using more Google products. This enables them to show results they know are not going to alienate the majority of users and give them real ammo to use to jack the price per click AND know when and who to show what ad to get the highest likelihood of return using personalization.

Personalization is the closest they've come to actual relevancy yet. It smacks page rank like a microwave beating a magnifying glass in the sunlight. The more they know about you the more likely they know which ad to show who when.

So, we want brands on the first page but not ALL brands just most with a few sites actually devoted to the search term to make it look good so we know. The branded sites will get the lion's share of  the clicks giving us much more reliable data. So, tell the programmer to give brands a boost.

OK, so we're back to putting on our programmers hat and we've just been instructed to boost brands. So now it's time to turn those complex marketing concepts into 1's and zeros., Uh oh, how do we tell the algorithm how to identify a brand URL? 

We can look at age of domain, how many direct searches get typed into the toolbar, how long people stay on the site, how many people actually purchase AND HOW MANY PEOPLE TALK ABOUT IT!

That's right, PEOPLE TALKING and taking some action that can be spidered. THAT, coupled with the other signals,  is how you 1 and 0 a brand.. Keep in mind it has to have that social mix because all the other signals could be present with one good link bait campaign and stumbleupon gives u a prestige placement for a day. BUT no one is going to impersonate a lot of people saying a lot of things. At least not yet. 

Reverse engineering is a boring venture even for those people who like doing that sort of thing so I'm not going to give u a bunch of examples that you can go and find on your own. Instead I'm going to show you one of the things I saw that got me to thinking along these lines. I'll let you do your own digging and draw your own conclusion. So let us use just one example.

Accepting the benefits of having someone else's idea of what I want shoved at me, we already know we are not all going to see the same results. Geo targeting, and personalization will give us all a different set of results but it's ok for this illustration. I've already tested this with a lot of people in a lot of locations sitting on a lot of IP's. You'll see what you need to see.

Do a search for bedding sets.  You'll see 7-8 of the top 10 organic coming from what we could all pretty much agree are big brands. Plus You''ll see the obligatory beddingsets.com (or similar), and maybe one other one that actually make a lot more sense for that term than the pottery barn. Or Overstock.com or Amazon. When I think of bedding sets I know pottery barn is the first thing that comes to my mind!

As you look at those results you can see that sure Debenhams does carry sheets but they also carry pencils! Why is that more relevant than a site devoted solely to bedding?  More importantly how did they determine which brand goes above which other brand? And we're back to the programmer.

If you'll go to the one or two EMDs, click thru and what you'll see is either no social presence or a small following,. NOW, go to about half of those big brand sites and every single one has f/b and twitter links plastered on the page. Now click thru and go to their facebook page and invariably you'll see at the top of their page, (if the page isn't completely covered with a pop up doing it), a large coupon offer, or some other enticement and it will say something like : LIKE this before proceeding. Like this to register for a free blah blah Blah. Now look at how many likes each off those brand sites has. Hundreds of thousands! if not millions!

EUREKA!

A spider can easily grab that like data and send it to the algorithm who can now cipher in that signal into the other over200+ (yeah right !), signals Google claims to use to algorithmically define relevancy and select the best result to improve the searchers experience!

There u have it kids. Want to be a brand? Get some 700,000 likes!

Peach Y'all

Bob Massa, CEO Search,King, LLC

Proud Parent of Blikini - the Social Media Management Machine

Social Media Marketing and Management