The Truth

Likes vs. Links for SEO
A Case Study, Finally.

If you’ve read any­thing about SEO in the past few months, you’ve heard at least 25 gurus claim that “Likes are the New Links,” but where’s the proof? We decided to find out for our­selves and con­duct a con­trolled, data-backed experiment.

Are Face­book Likes the new cor­ner­stone of SEO, or are the gurus full of s#%t? Learn More–

Who Started the Stink?

Some­time around April 2010, guys like Rand Fishkin began inno­cently spec­u­lat­ing that Google was going to begin count­ing “social sig­nals,” i.e. Twit­ter Retweets and Face­book Likes, in their rank­ings algorithm.

*I refrain from the cliche link-spamming, because unlike email spam, most auto­mated link-building tools cre­ate links that are not adver­to­r­ial, decep­tive or pushed on users.

The idea behind “Likes are the New Links” is that auto­mated link build­ing tools* are too advanced and the only way to com­bat their effec­tive­ness is to turn to the rel­a­tive purity of social net­works. In the­ory, peo­ple don’t Like or Share low-quality con­tent with their peers because they don’t want to be “defriended.” By this logic, “Likes” are a bet­ter indi­ca­tor of qual­ity than links. In real­ity, this is far from the case.

Over the fol­low­ing months, all of the usual sus­pects had some­thing to say about social sig­nals. The SEO “yes-men” fol­lowed soon after. To this day Sphinn is com­pletely drenched in “Likes are the New Links” stories.

How­ever, the tone of these sto­ries shifted dra­mat­i­cally from “here’s an idea” to “this is the way it is!” when Bruce Clay, owner of the inter­na­tional goliath search agency, Bruce Clay Inc.,  loudly announced that “Likes are the New Links” on his blog and at numer­ous con­fer­ences. It helps that Bruce is a board mem­ber of SEMPO and has tenure at every major search con­fer­ence (SES, SMX, Pub­con, you name it, he’s there), so when he wants you to know some­thing about search, he makes sure that you do. On the heels of Bruce’s loud and con­fi­dent asser­tion that “Likes are the New Links,” the whole indus­try burst into a roar of agreement.

With a blog post, a con­fer­ence speech, and a few online videos, Bruce trans­formed “Likes are the New Links” from a spec­u­la­tion to a “Fact.”

Sounds Good, but Where’s the Data?

With SEO being a global indus­try con­sist­ing of thou­sands, if not mil­lions of prac­ti­tion­ers in 2010, how could there be so much hype about Likes vs Links with­out an ounce of data to back it up?

• Had any­one per­formed a con­clu­sive, con­trolled study?
• Had any­one per­formed a regres­sive analy­sis on avail­able data?
• Had any­one even con­ducted an anec­do­tal case study?

Not that I could find.

Source: Danny Sul­li­van for Search Engine Land

Besides Bing selec­tively dis­play­ing Face­book con­tent on its SERP pages, and Google intro­duc­ing Real Time search, for search­ing Twit­ter and other microblog­ging sites, there was absolutely noth­ing to sub­stan­ti­ate the claims of Bruce Clay and his echo-chamber.

In fact, Aaron Wall, amongst other more sober SEO ana­lysts, says that “Social Rank” is likely on its way out before it’s on its way in. To sum­ma­rize his point, social net­works are just as eas­ily sus­cep­ti­ble to aggres­sive manip­u­la­tion as the link graph, if not moreso. What’s worse, it’s prob­a­bly harder for any algo­rithm to deter­mine what’s legit­i­mate and what’s not in the social world. After all, a 19 year old gamer con­tin­u­ously tweet­ing alter­nat­ing gam­ing rants and 4chan quotes and a bot doing the same would be hard for a human to dis­tin­guish, much less an algorithm.

Myth­busters: The Likes vs. Links Case Study is Born

Since nobody else was going to do it, we’ve taken the “Likes vs Links” ques­tion into our own hands.

The pur­pose of our case study is to 1) deter­mine if Face­book Likes have any impact on search rank­ings, 2) if so, how much and 3) if com­bin­ing likes and links is more effec­tive than either one alone.

Our hypoth­e­sis is that “Likes” will have min­i­mal or no impact on search rank­ings. We believe this because, cur­rently, we have no rea­son to believe otherwise.

Below is the exact method­ol­ogy of our case study.

(Note: We’ve done our best to con­trol all con­found­ing vari­ables. If you see any flaws or have advice for our exper­i­ment, feel free to put it in the com­ments below. We’ll do our best to imple­ment it. The case study will run over the course of 4 weeks, so that we do not pass any judge­ments pre­ma­turely.)

1. Find a Niche - We went on Google Trends and found a trend­ing key­word for a niche health prod­uct, the kind of thing you’d find tucked way in the back of GNC with a name suit­able for the next Harry Pot­ter title. We wanted some­thing low com­pe­ti­tion so that even if the impact was weak, it could still be measured.
2. Buy the Domains - Per­son­ally, I don’t think Google prefers .com, .net, or .orgs for rank­ings, but some peo­ple believe oth­er­wise. There­fore, instead of buy­ing keyword.com, .net and .org, my part­ner, Dax, opted to buy keyword1.com, keyword2.com, and keyword3.com. This way each has the same TLD, and the num­ber in the domain name affects each one sim­i­larly. We’ll release the domain names at the end of the case study so that read­ers can’t mess with the results during.
3. Put the Con­tent on the Pages - We cre­ated 1 spun arti­cle opti­mized for our key­word and out­putted 3 vari­a­tions to make sure the con­tent qual­ity, con­tex­tu­ally rel­e­vant terms and key­word den­sity were sim­i­lar. For the unini­ti­ated, when I say “spun,” I mean that we rewrote each sen­tence 3 times and ran­domly com­bined them to cre­ate unique arti­cles (Shame­less plug: this is some­thing we do at The HOTH link build­ing ser­vice). Each site got a stan­dard Word­Press install, using the default Word­Press theme with mild cus­tomiza­tion and an opti­mized pic­ture in the arti­cle. <title> and <h1> tags were optimized.
4. Day 1: Get Likes - On day 1, we went on a pop­u­lar crowd-sourcing site and paid ~175 peo­ple to “Like” sites #1 and #3. These are real peo­ple with Face­book accounts, not pro­fes­sional “Like ven­dors.” At this point we begin doc­u­ment­ing the progress of all 3 sites in the SERPs, includ­ing index­ing and rank­ing for the exact match term.
5. Day 15: Get Links - On day 15, we do 1 arti­cle sub­mis­sion for sites #2 and #3 (Shame­less Plug #2: This is exactly how we do The HOTH Link Vol­ume Mod­ule). We sub­mit the arti­cle to enough direc­to­ries to get ~175 arti­cle accep­tances and links back to the respec­tive sites.
6. Day 29: Mea­sure! Time to share what we’ve found. Make sure to sign up for our RSS Feed or Updates by Email to catch the find­ings of our case study.

There was a poll here, but it just wasn’t coop­er­at­ing. Instead, tell us in the com­ments below!

What do You Think?

Sign Up to Receive the Results of Our Likes vs. Links Case Study by Email or RSS.