Killing Google PageRank: Making Relevancy Irrelevant

This is old news, but a while ago on TW I posted that UPI, a 100 year old company, was overtly selling PageRank, even mentioning PageRank on their advertisement pages. Search works so well because they measure relevancy using things that are hard to manipulate or things that people wouldn't generally think to manipulate. Thus, if a 100 year old slow moving company is doing something you know that the method of relevancy they aim to manipulate is generally likely already dead.

Google will likely filter out overt link buys like this
Link Spam.
especially when they are marketed this aggressively on Google's own ad network
Buy PageRank from UPI.
If a link buy is so overt that people would talk about it, then an engineer or algorithm has probably caught it already. But that sort of example can be seen as a proxy for the market as a whole, and Google have also significantly lowered the weighting on raw PageRank scores over the past few years, because too many people know about it and manipulate it. Just looking at PageRank is nearly as useless as a meta keywords tag.

Directory Archives Update

I have sorta let the Directory Archives site go to crap (ie: be poorly maintained) for a number of reasons. I still like many of the better directories (like Yahoo!, DMOZ, Business.com, MSN Small Business Directory, BOTW, Gimpsy, JoeAnt), but outside of the top few general directories and a few high quality niche specific directories most directories probably do not pack much of a punch at manipulating Google's search results or delivering direct traffic. Given the flood of low quality directories, and their lesser value in manipulating search, I have not put as much emphasis on my directory of directories. All the following reasons play part in its reduced priority status:

  • Most directories are of limited quality. Building a high quality one is an expensive and time consuming process, and most of the people who have been creating directories over the last few years have not been concerned with quality. Most of them have been selling hollow PageRank to naive webmasters.

  • Search engines (especially Google) have limited the effectiveness of many low quality links, in some cases not only deweighting the value of low quality links, but if you have too high of a junk to quality link ration they may throw ranking penalties on your site or reduce your crawl depth or crawl priority. The reduced crawl depth and crawl priority (based on low quality inlinks and outlinks and duplicate content issues) also hit many of the directories themselves, causing many of their pages to be deindexed.
  • Social sites and consumer generated media add much more context to links than most directories do.
  • Social bookmarking sites have limited editorial costs and a huge number of editors, thus they are more comprehensive than most directories are, while having much higher profit margins.
  • So many people are blogging that if you can create something legitimately useful and get it a bit of exposure via your own blog or via the social sites then it is bound to pick up many high quality links.
  • Improving search relevancy coupled with this additional content makes directories less necessary. Various niche blogs, the Wikipedia, and other authoritative social sites have largely replaced most directories in the search results, thus reducing the direct traffic most directories send to listed websites.
  • Using the social sites is often a cheaper and more effective way of building a natural and diverse high quality backlink profile than by trying to build links from some of the lower quality directories. The social sites often lead to many secondary citations.

Rather than just killing off Directory Archives I added a social news and bookmarking section to it, such that it can still be used to help webmasters acquire good links and market their websites. In addition to listing social bookmarking and news sites it also has categories for sites like Squidoo and Work.com.

Teach Bots Language & Bottom Feeding Links

I recently added a brief overview of SEO to Work.com. Work.com is sort of like a Squidoo.com for business. An easy link from an authoritative site, and a chance to build co-citation. :)

As more and more companies push for consumer generated content there are going to be more and more bottom feeder link opportunities. In niche markets, if you use these types of resources, get a few directory links, and get a few links from relevant websites you are in the game cheap. If you think of search engines as ad networks trying to learn topic specific language sets you can help work yourself into the vocabulary by doing things like:

  • making sure your brand is mentioned on these types of sites

  • creating a bit of news
  • getting references on a few industry related sites
  • submitting to the correct category in a couple trusted directories
  • participating in a few forums
  • posting a good amount of content to your own site that references important industry related brands, organizations, and terms

If you are part of the language that defines an industry you have a distinct advantage over anyone who is not.

Using Keywords in Official Names

A brand can still push their main brand name (say Paypal, for example) while promoting their name as being Paypal Payment Solutions. Place more emphasis on your core brand name, but also make relevant keywords look like they are part of the legitimate official name to get a bit more friendly anchor text.

Nearly Every Successful Marketer is a Spammer

A9 recently largely died off. Why? Because few people talked about it, and it never gained any real traction. Google, on the other hand, even has people talking about their ads. Rand recently noted that he doesn't believe it pays to game Digg, while Loren notes that people are willing to game it for you for as little as $20.

Quoting Rand:

If you game Digg, you get none of these benefits. The visitors might click, they might even start reading, but if you don't have truly exceptional content, you're spinning your wheels - no one is going to remember you or your site as being anything other than a waste of their time; that's not a positive brand association.

...

If the content wasn't good enough to make the top of the link sites naturally, there's little hope that anyone who manages their own content will link to it.

Rule #1: If people enjoy it and vote for it then it is not spam.
I have seen average (or slightly below average) content become remarkable and Diggworthy through appropriate formatting and heavy use of Instant Messenger to seed the idea. If you can get the idea half way to the homepage (about 15 - 20 votes) before the general Diggers start voting you stand a good chance of making the home page.

Rule #2: Exposure leads to more exposure.
And in spite of the fact that you will not get many links to cheesy content, I have seen somewhat cheesy content garner high trust links from old school media sources. You really only need one of those types of links for the Digg spam to pay for itself. And those links are going to be hard for competitors to replicate (unless they know how you did it).

Is it wrong to pay people for exposure? If so then why do search engines teach content publishers to blend their ads into the content?

Rule #3: Most members of the media are overworked or lazy.
Another thing to think about, is that many of the people at mainstream media sources are lazy, underpaid, or overworked cogs. In the same way that some journalists have swiped ideas from bloggers before, many of these journalists may rely on these social news sites to find new things to mention or link at. And I have seen anchors at one social news site submit one of my stories to their site only because they found it on another one.

Read more about the relationship between public relations and the media.

Rule #4: New typically means easy to spam.
The newest systems are generally going to be some of the easiest to spam.

And the spam doesn't look or feel or smell like spam if people are reading about themselves. SEO Blackhat wrote a funny article about how to get seen on Digg and over 4,000 people voted for it.

Rule #5: Older systems are typically more expensive or harder to spam.
A year or two ago it was far easier to spam Google using mini domains and keyword rich anchor text, but since then the algorithms have been placing more and more weight on citation based authority. Newspaper sites are reporting a large increase in online exposure. The publish the same old bland content and are competing with more and more sites. I don't think I would be in error to assume that a large portion of that increase is due to bias shifts in Google's (and other engines) algorithms.

Rule #6: To be successful, you have to be a bastard to somebody.
In an interview with Rolling Stone (available free via iTunes), John Lennon stated that you don't get to make it big without being a bastard. And The Beatles didn't get to become The Beatles without being serious bastards.

Rule #7: Almost everybody spams.
Any for profit system has rules set up that help it make money at the expense of others. If you are starting from nowhere you really don't have much to lose by being a bit aggressive. After you establish a strong brand then overt spamming may not look as appealing on your risk to reward ratio scale, but off the start it shouldn't hurt to be a bit aggressive, and if people are going to hold that against you forever, then screw em.

You have to spam somebody to get people to grant you enough authority to influence other markets. After you gain enough influence you keep pushing after other markets:

Google's Eric Schmidt predicted that "truth predictor" software would, within five years, "hold politicians to account." People would be able to use programmes to check seemingly factual statements against historical data to see to see if they were correct.

A statement like that makes you wonder if a Google ad campaign might help determine what truth is perceived to be.

Even after you have lost touch with your core purpose businesses keep pushing for growth:

TRUSTe’s Fact Sheet (2006) reports only two certifications revoked in TRUSTe’s ten-year history... According to TRUSTe’s posted data, users continue to submit hundreds of complaints each month. But of the 3,416 complaints received since January 2003, TRUSTe concluded that not a single one required any change to any member’s operations, privacy statement, or privacy practices, nor did any complaint require any revocation or on-site audit.

TRUSTe has only a small staff, with little obvious ability to detect violations of its rules. Rule violations at TRUSTe member sites have repeatedly been uncovered by independent third parties, not by TRUSTe itself.

Is there a single profitable well known online business that doesn't spam or at least pay others to spam for them? And, at some point, did they spam to get where they are?

I linked to this before, but I love this audio file.

Here a down and out. There a game fighter who will die fighting.

Exact Match Domain Names in Google

Search relevancy algorithms are ever changing, but I recently snagged a good example of Google placing significant weight on exact matching domain names. When you search Google for search engine history there are over 20,000 exact phrase match pages and over 90,000,000 matching pages. The #6 result in this screenshot is SearchEngineHistory.com, which is a site that I never really developed. It has no inbound links on Yahoo!, Google, or MSN (as you can see on this screenshot and that one). Also worth noting that SearchEngineHistory.com is a single page site, and with NO link authority it outranks a LifeHacker post that has the exact matching phrase in a page title (and LifeHacker is an extremely authoritative site).

Why could Google trust domain matches so much? Because they are often associated with brands which protect their trademarks more vigilantly than in the past, and there are so many domainers and so much vc money placing premiums on domain names. To get an exact matching domain it is probably going to cost you something (either lots of money or the foresight to be an early believer in a new field), so that in and of itself is some sign of quality. For example, today I tried buying a non-word 5 letter domain for $1,000 and the domainer turned me down stating that he turned down 5x that much last week. About 3 years ago SeoBook.com cost $8, largely because the standard frame of thought in the SEO market was that there was no market for a book or ebook.

Andy Hagans on Quality Content?

Andy Hagans is advocating quality content AND advocating it within a quality content post.

What is the world coming too? Somebody check the phase of the moon!

Old Gold

If you write hundreds and hundreds of pages about a topic odds are that eventually one of them is going to rank, get some decent self reinforcing links, and then keep ranking. This is especially true if you are writing about a modern technology or a field that is rapidly changing. One of the reasons exposure on sites like Digg and Del.icio.us is so valuable is that it earns you unrequested secondary and tertiary (and fortuary, hey wait, is that a word) organic citations. Some underfunded mainstream media sites just link to whatever ends up on sites like Del.icio.us that day, then other people find those channels and link to you from there as well. It is equally cool and lame, but perhaps a bit more cool if you are on the receiving end of the linkstream.

Longterm the key to doing well on the web is to do things that are strong enough that they build unrequested links.

So what if you are already ranking #1 for a keyword on an old page? Is it ever worth editing it?

I have a page which got about a half dozen unrequested .edu links back in 2004. The page was probably of average quality, but easy to cite, because it looked comprehensive. As time passed I added a bit of info to the page here and there but did not go through to format and edit it...those changes, coupled with rapid changes in that field meant that the page went from average to below average quickly.

It still ranked #1, but that page has not got a single .edu citation since 2004. What if I would have made that page far better? Have I been throwing away a .edu link a week for the last year and a half? Likely. And it gets worse too, because as that page would have got cited it would have lead to secondary and tertiary (and fortuary, hey wait, is that a word) organic citations from people who were passionate about that topic.

And had that site gained another 50 or 100 .edu links it would have doubled or tripled the value of that site. The authority from that one page would have carried that site.

Now I am not a fan of going through and editing everything over and over again, but if you have a couple core pages which capture powerful ideas it is worth it to make those as good as they can be. And if you already rank, then you are just leaving links on the table if those pages are average. Clean them up a bit and get the love you deserve, you obviously deserve it if you are already ranking :)

Some marketing is push. Other marketing is pull. What makes SEO great, is that when you figure out what ideas to target your pull marketing is self reinforcing while others are pushing pushing pushing and never able to catch up.

Digging for Links

When you are doing SEO you want titles that are rather directly informative...you need to be descriptive. But that is not how you promote a linkbait.

Too much sensationalism causes you to lose credibility, but if you are starting with none then you might not have much to lose by testing different things. Take this post, for example. Let's analyze it.

  • The person who wrote the story about Google submitted it to the Apple category.

  • The post is half-assed research, passes opinion as fact, and is completely wrong in it's conclusions, but
  • The post is titled Google's dirty little secret

Thus despite multiple layers of ineptness it is passed off as good information based on the title alone. It made the Digg homepage. A good linkbait starts with a good title.

Need help with your headlines? Go get some magnetic headline love.

Another tip is that for the amount of effort you need to put into making a piece of information, you are typically going to get much more out of it by making it biased than by aiming for vanilla. Your bias is what people subscribe to, want to believe in, or want to discredit.

For example: I think Iraq for Sale is a film every American should see, largely because those who allegedly support a free market system think that the uncontested multi-billion dollar government contracts full of fraud sent to scumbag corporations are an acceptible business practice. And they only get away with it because people argue on the rhetorical or idealistic levels instead of talking about what is actually happening, and the media is generally not honest enough to report some of the news themselves...they are too tied to profit to allow themselves to.

I bet someone comments about that last paragraph ;) Also notice how I lined out I think. If you want to be controvercial an added way to do it is to present opinion as fact, but be forewarned that if you go to far with that it makes you an easier lawsuit target. But if you plan it out correctly lawsuits can go right into the marketing budget. With some stuff it almost seems like that is how Google does it :)

Linkbaiting is all about emotional reaction or being memorable...that is what leads to comments and citations.

So how else do you make your story comment worthy or citation worthy? You build up a following over time. Those who read your site may Digg, Netscape, or Del.icio.us your posts.

How else do you do it? Have an instant messenger list a mile long, and email. Beg your friends. Time your post, bookmark your site, and then light up contacts via email and instant messenger.

There is not a lot to the linkbait formula

  • time your post for a launch when your friends will be around, use a catch title, Digg your own post

  • be really biases, or format your information so that it LOOKS fairly comprehensive
  • beg friends
  • make it easy to bookmark your page by placing the following type of code on your site

{ Post to del.icio.us | Post to Reddit }

Note: if I was trying to get this page to Digg's homepage I probably would have used something like Digg is Too Easy to Game: Here's How as my page title.

Two Pages = Double Listing Love

If you want to make a site that looks legitimate and is well structured you probably only want to have one main page for each topic, with sub pages working to further expose sub-topics. But what do you do if you are tracking your results and are making a thousand a month or more from a single page? Some algorithms are somewhat literal in nature, while others are more elegant and look for more natural writing. I am still trying to tweak a page in to rank for all varieties of it's core main phrase in Google, but in the course of tweaking it in to match Google (by making it more elegant and less literal) that page does not score as being as relevant as it once was in Microsoft.

The ideal solution would be to just keep getting authoritative links until that page was viewed as the ultimately authoritative topical resource by all major engines, but unless you have real topical authority and high quality content it is going to be hard to get legitimate citations on the conversion oriented page. And getting it low quality link spam is not going to be the most cost efficient method if I care about the long-term health of the site. In fact, without trying to get any spammy links the page picked up hundreds of them just by ranking well.

What is another solution? Use the main page with the most link equity to target Google since they have the largest search market share, but also consider creating a second page on the topic which is more literal in nature. The second page can be about the history of the topic, background information, future of the topic, how that topic fits a location or a minority, saving money with topic, do it yourself with topic, or frequently asked questions related to the topic.

After making the first page less literal and creating the second page that was exceptionally literal I checked back on the rankings of both pages for some of the core keywords. For many of the phrases those pages targeted I scored a double listing in both MSN and Google.

As search engines change their relevancy scoring algorithms they may not only change weather they match deeper or shallower pages, but they may also change what they are willing to rank based on how literally it matches the query.

Another way to look at this phenomena is that if you are working from a new untrusted domain and are trying to create the backfill catalog of content for your site, it may make sense to make some of the early writing more literal in nature since it will be easy to rank well in MSN for it. As you learn more about your topic, get a bit of a following, and have some topical authority it may make sense to go back over some of your most important content areas and create new pages which are less literal in nature.

Pages