2013-12-27

Rap Genius “taken off the internet” is a post from the best Link Building Tools available worldwide.



Google Penalizes popular lyrics site on Xmas-Morning for Spammy Linking Tactics.

So it’s Christmas Time again, and Google did it again. Lyrics site “removed from the internet”.

Following a notification on YC and Twitter and roughly 24 hours due diligence by Matt Cutts himself RapGenius was penalized by Google for a link scheme.

So they say.

@mmasnick @RapGenius we're aware and looking into it.

— Matt Cutts (@mattcutts) December 24, 2013

To quote Matt Cutts:
“The spam fighting never stops, even on Christmas.“

@joehall the spam fighting never stops, even on Christmas. Hope you're having a good one.

— Matt Cutts (@mattcutts) December 25, 2013

What happened to Rap Genius?

RapGenius was pushed to page 5 or 6 for basically every search phrase one typed into Google.



Rap Genius disappeared

This disappearing of RapGenius also made a lot of users believe they were “taken off the internet” completely, which was essentially the case from a normal users’ view but probably also investor’s perspective.

NOOOOO @RapGenius GOT DELETED CHRISTMAS IS CANCELLED DAY RUINED

— James Reese (@ReesesPiece94) December 25, 2013

Rapgénius existe plus ? :0

— Πatteø (@Mattious_) December 25, 2013

Il ont coulé ou quoi Rap Genius, j'trouve plus leurs site

— Ilyass Moussa (@Pouss_Mouss) December 25, 2013

Double Traffic Drop -60% and -52%

The search traffic of RapGenius dropped 60% the first day, then another 52%

This means a drop from an already weak 1.2 million unique visits on Dec-24 to 493k uniques per day and an even worse 234k on Dec 26 according to quantcast.



WOW. That’s a sharp drop – even or especially for VC backed companies that celebrate their success, community and technology (a special way to add annotations on the web-page, much like the yellow bubble comment function you hate so much in Word, but hey it’s something you don’t have on every website)

The Link Scheme that made RapGenius disappear?

@mmasnick @RapGenius we're aware and looking into it.

— Matt Cutts (@mattcutts) December 24, 2013

The scheme was asking bloggers to link out to a dozen or two deep pages to Justin Bieber lyrics (not exactly Rap) in exchange for a “Wohoo tweet” from them.

All that got out in public because a guy called John Marbach outed the tactic itself on his blog, after simply asking for what that “affiliate program” was.

Of course that caused some confusion pretty fast

Growth hacking = using 2005-era spam tactics. http://t.co/5ISCPmMEkp cc @samfbiddle @nitashatiku

— Max Woolf (@minimaxir) December 23, 2013

My take on this whole “blackhat tactic”

Asking people especially your fans to link to you is not against Google Guidelines

Incentivizing that with a tweet is well… probably “paying” for it in Google eyes now, but it’s VERY common for business to encourage their fans to link to them, and if not their fans, who would be better link sources

This example was MAYBE seen as unnatural approach by Google

This example was MAYBE just the tip of an iceberg that got this whole story running for Google to do a deep-dive.

But hey, that’s not so Blackhat SEO as some people tend to call it.

What’s even more special about this approach is that they asked to have a dozen links included, all to the same site, which is again not very SEOish… SEOs these days try to camouflage their linking efforts a lot better.

If you are a rap-fan and you have no problem putting up a dozen links to Bieber lyrics, which in my book is not rap, then it’s your own take. If you want the whole world to know that you also like Bieber and RapGenius supports that – why not?

Google forced to act

The fact that RapGenius is VC backed, and indirectly even via Google and even Matt Cutts himself and the whole thing gotten so public put quite a lot of pressure on Google to act.

If GOOG won't act on vc-backed brand webspam for fear of harm to GOOG Ventures, then #sec and #doj should investigate http://t.co/VoJsxOEUEZ

— Cygnus SEO (@CygnusSEO) December 24, 2013

But if Google hurts a VC backed company they might start to get blackballed for their own Google ventures. The flip side is if they don’t penalize, then they open themselves up to lawsuits where discrimination clearly takes place. Not an easy decision here.

The problem of being too loud, crying out “have won then SEO game” and pretty aggressive foul mouthed press coverage like discussed in this post “Spam links and the dangers of not giving a fuck” probably didn’t help.

Essentially Rap Genius made Google look bad, and you don’t want to do that if you base your business on free traffic from Google.

Basically all these actions forced Google to take a deeper look.

Based on this history the whole story spread for them being penalized for this “pitching links for a tweet”, but Google has never confirmed that it was ONLY that specific thing they got penalized for.

And I don’t buy that cake either.

Pitching Links for a Tweet for a penalty? Or is there more?

Now that “Yoo waddup please link to me mail” is certainly something Google didn’t like, especially not with the incentive to get a high traffic tweet, but from my point of view there’s more to the story.

We’ll try to find out.

If not today, then in the very near future.

RG already outed a couple links they think the competition bought or acquired in an unnatural way.

Rap Genius says sorry

RapGenius founders now published a “mea-culpa we did what everyone did” post on their site and even outed some of their competitors, namely

AZLyrics.com

Metrolyrics.com

Lyricsfreak.com

Lyricsmode.com

Lyrics007.com

Songlyrics.com

And they very rightfully say that the lyrics space is a pretty spammy one, their competitors were doing the same or worse, and that Google should take a closer look at it.

They even went on to classify their competitors’ strategy according to 4 different areas of link rules in the webmaster guidelines. Nicely made, and I’m surprised how fast they out all that together. Or was it all prepared already?

I think it will def. be worth looking into their competitors and the space in general, given what they claim and what I’ve seen in a quick review.

I was quite surprised to read only a few hours later on Techcrunch that Google and Rap Genius “are working on a mutual solution” – sounds a bit exaggerated to me. What is probably the case in our technical terms is that they got a manual penalty in their Google Webmaster Tools and prepare to file a reconsideration request.

About the “Google Victim” Rap Genius

If you feel just like this guy

Never having heard of Rap genius before now I feel little sympathy for them.

— SEO Theory (@seo_theory) December 26, 2013

Here’s a quick roundup on RapGenius.

RapGenius.com is the new-kid-on-the-block of lyrics sites, and instead of just dumping a 99$ database from some SEO forum onto a domain to capture on AdSense traffic, they created a mixture of Wikipedia, Urban Dictionary and well, a lyrics site. They even secured $15 million in funding about a year ago and are even funded by YCombinator now.

Those $15 million were for the fancy feature of being able to have a community annotate the lyrics source on the fly.

About the “Lyrics space”

A quick overview of how SEO and link building works in the “lyrics” space and some special attributes we see in that space

VERY long tail content

VERY long tail in link anchor texts…
for example “damn it feels good to be a gangsta” is a money keyword phrase (with that a lot of sites, including Youtube compete for with very similar content – either the video, the tune or the lyrics.

Links earned or built again can have a huge variation of anchor text for a single page. In the given example above, it could be any part of the song title, including the “Rap Genius” source (a compound keyword then) or just a brand keyword phrase like “via RapGenius.com”

A LOT of deep links. The deep link ratio for Rap Genius is over 91%

Large content sites like Lyrics sites are targets of scrapers and “real black hats” from the Pharma space for example also, which makes a lot of those links not desirable at all and should in fact be disavowed pro-actively in terms of what we call link risk management.
See below for more details

A quick look at the backlinks of Rap Genius

Pharma spam links

If you look at those sites linking with above anchor-text like “damn it feels good to be a gangsta” you find e.g. these here

And one of them at http://www.telfax.pl/wp-content/themes/dacota/MTA1.php

looks like this- to GoogleBot. Only. Notice that lyrics link down there that was probably generated like the rest of the post that seems to target the medication “Cymbalata 30 mg”

Of course visited thru a normal browser it looks very different and redirects the user to a site where you can actually shop for that medication.

This redirection is a typical cloaking scheme that SEOs and Google consider black hat, and it probably also involves hacking or exploitation of vulnerabilities on those sites where the links are posted. So the owners giving those links probably don’t even know about it.

Each of those hacked pages have a couple or dozen links to them themselves, probably spammed by the RX/hacker/black hats themselves, not Rap Genius. But essentially it’s creating a large link network. And if we can find it so fast, image what Google sees.

We found 44 of such toxic links to that single subpage with that keyword phrase alone. That doesn’t look too good.

If you’re curious, the links going to these cloaked low quality pages linking to RapGenius are even worse, chines guestbook spam.

Link Velocity Trends Spike

Looking at the Link Velocity trends, one training LRT Superhero can immediately see how their link growth spiked up like crazy just in November. Viewed through a Google’s Spam engineer that chart looks pretty unnatural. Frankly could be just the result of that new affiliate program.

IP and Domain Diversity

A Quick Backlinks Analysis shows a huge number of link counts from the same IPs, and usually this is an indicator for sitewide links, which again are often bought.

Drilling down per domain we can confirm this

No Paid Links?

High PR “Blogroll links” on a couple 1000 pages like http://djservicepack.com/look pretty artificial on the one hand.

But others are just countless deeplinks in “top 5 posts” about the lyrics of the week on TheGrio.com that I wouldn’t call bad per-se. Don’t get me wrong, maybe I’m just not the target market for this, but given how many top-lists of some stuff get published in the SEO space, those look legit to me, and it’s an MSNBC property, so would expect some editorial discretion there.

On the other hand, as they point out, some of their competitors have links from pages that specifically spell out that the links placed are “direct links for PR reason”. PR being Page Rank, not public relations.

It would be good for Rap Genius to NOT have any such obviously paid links to their site – further review and inspection pending.

Overdone Power vs. Trust?

Yes – like many other over-SEOed sites, we see a slightly higher Power than Trust indicator here. We also see that the site is hosted in CR (Costa Rica). Two reasons come to mind – intellectual property and taxes. But that’s another story.

We also see that the number of backlinks to that (pretty young) domain is more than 3 times the next biggest competitor AZlyrics. Now if you have $15MM to spend, you can sure get some links for that, no question, but I doubt they are all bought.

Super-Long-Tail Profile

As mentioned earlier, what’s really interesting in general in the Lyrics space is that both, the content and the link profiles are super-long tail. That’s almost the perfect anchor text variation you can get by everyone and his dog linking to some song with his favorite quote of the song. Compared to all other niches I know this is the niche with the widest range of anchor text variations possible. Quote-Database sites maybe come close.

This makes it harder – a lot harder – for Google to detect Anchor Text over-optimization.

Justin Bieber Rock vs. Rap Genius subdomain

What’s notable is that this “affiliate scheme” was setup for the most recent and widely popular Justin Bieber songs, asking for links to the “rock” subdomain. This is actually another weakness here. While the site is widely popular for Rap, the branch out into other high traffic areas like Bieber justify that sub-domain, but on the other hand make it a lot more vulnerable as Google treats subdomains differently.

And guess what – with that sub-domain having substantially less links they have a wider skew between the strength of links (they probably got with this promo) and their trust, compared to the main domain

As you can see below, that CEMPER Power*Trust level distribution is the poorest for the “rock” subdomain, which the overall main domain has the best, strongest and trusted links.

Anchor Text Breakdown

Looking at the anchor text breakdown for the rock.rapgenius.com domain alone, it becomes very clear that A LOT of money keyword phrases were linked. What’s interesting is, that many of them were augmented with the brand name, which we call “Compound keyword class”. This is usually a very seldom keyword type. But in this case it’s the most dominant one.

An interesting question that pops to mind is, if Google did actually tweak their variables for those usually “safe” “compound keyword phrases” to a higher risk level, given how many of them were used here.

The overall risk calculated, thanks to less riskier compound keywords looks quite ok, and this subdomain would not trip a Google Penguin filter by our standards today. So far.

However… let’s see if Google might change that.

Experiment Compound vs. Money ignore

Now if we were to say that Google would call those super long money keyword phrases too long and ignore that the “rap genius” brand is mentioned too, thereby NOT counting them as compound keyword, I think the overall story would look different already.

A quick re-classification to make those count as “money”, free reprocess and here we go.

You see after “correcting” that special case of only 2 phrases on thousands of pages with keywords from compound to money keyword phrases, the picture looks a lot more “natural”.

The overall risk went up for that domain, but it’s not yet at that magic 1000 edge where typical Penguin Victims (or higher) sit.

They still have a ton of moderate risk links that I would consider for a review, everythin riskier of course as well.

That being said, not even the rock.rapgenius.com domain looks spammy enough for an algorithmic Google Penalty, which can be confirmed as only the manual action by Matt Cutts and his team himself led to the penalty we see.

Quick Competitive Landscape Analysis

With our Competitive Landscape Analyzer we quickly tried to verify the claim that “the others” are doing so much worse, and did a quick analysis.

Some findings are pretty straightforward

Judging by CEMPER Power*Trust the competition has a couple stronger and more trusted links than Rap Genius. Until now! Keep in mind they get major media coverage, so that will change VERY soon. Good for them.

The distribution of money vs. brand keywords is a lot better for Rap Genius. Their use of compound keywords helps here just like the general linkage with their brand name in general over the competition.

The percentage of in-content links is considerably higher than their competition. Nothing bad per-se, but worth a closer and deeper look.

The number of links from identified Google-Authors is more than double than their next competitor. Now we’re not strong believers that this is a helping factor right now, but it tells a story about their link profile.

Oh and A LOT more links from blogs than their competition. Now this again could confirm that their “affiliate program outreach” yielding more links than expected and potentially helped them already.

Conclusion

This is just a quick dive into what Rap Genius and their competitors did, in addition to sending out that specific email. There’s a lot more going on in that space in general and potentially also with Rap Genius, but I tend to agree with them that they are cleaner and more natural than many of their competitors in the niche. Simply also because those were spamming and jamming like everyone did in 2005 or so, and Google changed the rules on that.

It will be interesting to look at the Rap Genius case in specific and the whole lyrics space in general. I mean who keeps Google from just soaking up all those lyrics in their own database and integrating it with their services Youtube and G+ anyways? The music is there, the video is there and the lyrics are somewhat copyrighted… that wasn’t such a big issue for Youtube and is certainly not for Google in the past. I’m actually surprised Lyrics.Google.com doesn’t exist yet.

Despite this manual penalty that certainly hurt (heck from 1.2M to 200k in 2 days!), I think long term RapGenius have a bright and prosperous future ahead as long as the lyrics space remains independent products, i.e. websites depending on Google. Hu – was that a paradoxon?

All the media coverage that Rap Genius got over Xmas is just like that – a XMAS-present in highly trusted links from across the world. Those trust links outdo all the other link building efforts they did in the past.

This advantage will help them and hurt their competitors in the long run, so there’s fun ahead probably.

Google also got their PR stunt right again.

Like every year around Xmas, a low news frequency time, a popular name gets banned or penalized in Google and and the mainstream press starting from NYtimes, NBC, Techcrunch and of course niche sites and SEO blogs report about it.

Report Google’s story that you shouldn’t mess with them. Not even if you’re JC Penney, Overstock, Interflora or a VC backed hipster gang who rebuilt the lyrics site concept in a pretty cool “web 2.0” fashion.

The question that remains only touched a bit – how bad is Rap Genius really, compared to the rest of the industry?

If you would like us to dive deep into the song lyrics niche in the next days, let us know in the comments.

If you want to get notified about upcoming case-studies, you can sign up here.

Your First Name*

Your Work Email*

Show more