Like Me? Follow Me.
Over the last couple of years Google has been rolling out updates that aim to:
- promote legitimate brands over affiliates and doorway sites
- reward great content
- take user experience into account - including using reviews and social signals
- devalue or even punish sites using links to manipulate listings via networks, paid links, etc.
- penalise sites for aggressive optimisation, tactics such as keyword stuffing and unnatural anchor text
All of this is meant to help promote good businesses with good websites that contain good content and deliver a good service.
Except...in many instances, the exact opposite is happening with fly-by-night, keyword heavy, content-short sites ranking off the back of an exact-match domain and a burst of low-quality linkbuilding.
The Example SERP
We have a client (more than one, actually) who works in a particularly aggressive sector related to the legal services industry. For a couple of years they were doing their own online marketing, adding keyword rich content to the site and blog regularly. We did a little initial optimisation and training and some linkbuilding when the site launched, but they were pretty much doing everything themselves until about 8 months ago when they came back and asked us to put together a proper online marketing strategy for them.
The first thing we realised was that the sheer volume of content had quickly outgrown the information architecture of the site and the second was that in their desire to get lots of content they had inadvertently caused themselves a massive keyword cannibalisation issue - and had often written multiple pages on the same thing. Most of the pages were also massively overoptimised for a particular key word (which is also part of the brand name). It was an easy mistake to make, but we were concerned this was damaging them, and so we embarked on a mission to tidy everything up and make sure the site content was as good as it could be.
In the process of fixing the problems we broke the site - this was about 2 months before Penguin and the overoptimisation penalty. Nonetheless as we moved pages, redirected pages and changed the names of things we kept accidentally spawning 404 errors and finding duplication problems. The listings which had crawled over 6 weeks from outside the top 100 for many terms onto pages 3 and 4 suddenly dropped into oblivion.
We kicked our cleanup efforts into high gear and sure enough, within about 10 days of us finally fixing all the 404s, redirect chains, revising most of the content and removing a couple of links that looked particularly bad (despite being relevant), the site started climbing back up the search results. Then Penguin happened and suddenly what had been a fairly stable SERP went insane.
Some of the businesses that had been on page one for years - high profile businesses with massive offline budgets - disappeared and what appeared in their place was a rotating set of EMDs run by no clear business all with identikit layouts, giant data capture forms and very little in the way of meaningful content. The top 40 was changing almost daily (and listings are still shifting like an old school Google dance).
What strikes me most about the SERP I've analysed (some of the SERPs for other terms in this vertical don't have these issues, at least not to the same degree, many do) is how the top 30 seems to reflect everything Google was trying to destroy over the last 2 years. I've pulled some statistics across the top 30, leaving out results for large well-known informational sites such as Which and the BBC and put them into the table below to demonstrate some of the most striking problems with the result.
The metrics I've chosen are the most easily comparable I could think of:
- Number of referring domains
- Number of inbound links
- Number of referring C-class IPs (to demonstrate networks)
- Majestic SEO Trust flow - to demonstrate quality
- Majestic SEO Citation flow - to demonstrate link authority
- Domain Age - typically a trust metric (and to show how Google is no longer making sites prove their staying power)
- IP Address - to identify networks with multiple sites ranking in the top 30 which Google should be weeding out for QDD
- MozRank - another comparable metric to demonstrate the relative strength/weakness of the backlink profile
- Indexed Pages - If it's all about content surely Google should prefer sites with content over sites with nothing but a giant form and content 'borrowed' from Wikipedia
- Business Ownership - to try and suss out which sites are brands, which are just microsites set up by a company to try and dominate a SERP and which sites seem to have no brand/ownership at all
I haven't used any social metrics because a quick look at the top few sites didn't turn up any notable social mentions.
It's pretty bad out there folks
So here are some of my conclusions. One company has 4 sites in the top 30 - all easy to spot (even though they are separate C-Class IPs). One company has 4 sites on the same C-Class IP. There's also 2 other companies on IPs in almost the exact same location to that company. Another company has 3 sites in the top 30. SO that's 3 companies taking up 11 out of the top 30 rankings in a competitive SERP.
One of the companies with 3 sites in the top 30 has a total of 14 pages indexed. Another has a 1 page site. 7 of the top 30 sites are on domains registered in 2012, one as recently as 23rd April. Another 3 were registered in 2011.
There are 4 domains either registered anonymously or with no clear ownership established - including no company registration or VAT number displayed nor any links to or affiliation with the appropriate regulatory bodies.
Only ONE company in the top 30 doesn't have either word from the search phrase in the domain name. Only 5 appear to have brand names at all (which happen to contain one of the two words in the search phrase in the brand).
The link data is equally shady:
- the site at #1 has acquired 3,597 backlinks from 608 domains since it was registered in February of this year.
- the site at #2 has acquired 4,386 links from 671 domains since it was registered in December of last year.
- The site at #10 has acquired 1,231 backlinks from 326 domains since it was registered in April of this year.
- the site at #29 has the highest MozRank at 5.39, the highest TrustFlow at 50 and the highest Citation flow at 39 and has been around since 2007 (no, this is not our client)
An unscientific examination of the links pointing into the top sites show splogs and dozens of unused forum profiles - the very low value stuff Google insists shouldn't be working anymore.
Even a year ago if you built links as quickly as these sites, you'd be sandboxed straight away.
To be honest, to any SEO who has been following what Google have been claiming about their goals for the changes they're making should be able to draw conclusions about what's so shocking about this search result by looking at the table below. (Click to enlarge and open in a new window)
The sites ranking well seem to be taking a very cynical and short term approach, buying up EMDs that seem to rank well almost immediately no matter what sort of garbage goes on them or points to them, knowing that they may only rank for a little while, but attaching no established branding to them means the companies benefitting from the garbage SERPs won't have their brands damaged.
What is surprising is that with the evidence this obvious, Google seems unable to stop it algorithmically. How unable? Well, we bought an EMD in this niche; uploaded some vaguely spun content originally taken from Wikipedia; visited it while logged in to a Google account; and in under 48 hours it is ranking #12 for the term it matches.
Bad for Users, Bad for Businesses, Great for SEOs and Web Developers
What is most infuriating about this state of affairs is that it's enabling a particular sort of cottage industry to prosper at the expense of some good genuine businesses delivering good, genuine services:
- SEO or Web Developer/Designer buys dozens of EMDs in competitive niches, throws up thin site of 5-10 pages, mainly from scraped newsfeeds, articles, etc. (yep, we've seen dupe content on these sites ranking well).
- SEO/Developer gets a few hundred links from directory submissions, splogs, forum profiles and article submissions (yep, this is working still too).
- SEO/Developer suddenly outranks older, well-established websites for keywords in the EMDs, pushing them off page one.
- EMD site captures leads that would go to genuine businesses.
- SEO/Developer contacts businesses he's pushed off the first page and sells them back the leads they would have had.
- SEO/Developer, realising their tactic probably won't work long-term, then suggests to said businesses that they should simply purchase the site (or sets of sites) that are outranking them, and sells off the sites at inflated prices.
- Google adjusts and penalises said sites and businesses are out of pocket thrice over - for the drop in traffic, for the cost of leads, and for the cost of the websites that outranked them.
- SEO/Developer picks another set of keywords/niche and begins again.
I was going to write an incredibly sarcastic concluding paragraph here about the amazing awesomeness of the Penguin update, but I don't really think I need to because any SEO consultant who's gotten this far will already anticipate anything pithy I might say about how broken Google's results are in this sector.