$1,000 SEO Link Value Reward Contest

Jun 7, 2011 • 8:54 am | comments (31) by twitter Google+ | Filed Under Link Building Tips & SEO

link rewardMichael Martinez, a frequent commenter here, posted a $1,000 reward for someone in the SEO industry to provide a clear 100% positive proof with evidence that links work to rank sites in Google.

His rules are clear:

  1. A test that any SEO of any experience or competence can use
  2. A test that works 100% of the time.
  3. A test that proves to any reasonable person that an existing link IS or IS NOT passing value.
  4. A test that works with ALL links, not just links with unique anchor text.
  5. A test that can be performed in 10 minutes or less.

But are the rules achievable? Michael thinks not and many SEOs think not as well. And Michael is so confident, he is willing to put $1,000 of his own cash on the line to prove it (or disprove it).

You have until June 30, 2011 to prove it.

There is one forum thread discussing this contest, it is at Cre8asite Forums and I wanted to highlight one comment from administrator Ron Carnell:

Life. It's not always science, it's rarely clean or pretty, and it's never ever ever precise. But you know what else it's not? It's not always arbitrary and it's rarely as simple as pure luck. It's one choice leading to another choice leading to the next. To think otherwise is to ignore the fact there are invariably some people who do it better than others. Not just once in a while, but usually. Not every single time, but still more than most. Not scientifically, not precisely, but nonetheless ... reliably.

Up for the challenge? Details over here.

Forum discussion at Cre8asite Forums.

Previous story: Google Web Search Results May Add More Pictures



06/07/2011 01:12 pm

"A test that can be performed in 10 minutes or less." This is hardly scientific. Can a drug maker prove results in 10 minutes? ROFL. 

Jeff Yablon

06/07/2011 02:51 pm

Wow. Can you say "bullshit"? These two criteria conflict: +A test that works 100% of the time.+A test that proves to any reasonable person that an existing link IS or IS NOT passing value.   100% efficacy makes the "reasonable" think not meaningful. Conversely, "Reasonable" means you are actually thinking and don't need "proof" Sad. And he got you to fall for it, which is even sadder Jeff Yablon President & CEO Answer Guy and Virtual VIP Computer Support, Business Change Coaching and SEO Consulting/Search Engine Optimization Services

Chris Boggs

06/07/2011 03:01 pm

lol is this link below your sig part of the test? :p

Barry Schwartz

06/07/2011 03:09 pm

FYI, you can reply to comments via email. Not sure if this is what he did. :)


06/07/2011 03:29 pm

dunno about such a test and it's validity...but, as others may have noted, this "contest" idea sure gets some serp results and SEO rep, eh! nice one Michael! :-)Jim

Rank Chef

06/07/2011 03:35 pm

I'm going to do somebody a favour and let them win $1,000 on my behalf  :) The test is very simple (and I've done it before to test on-page ranking factors): 1. Host 2 files on your website with the same content. 2. Include a unique string inside the file that will be easy to search for in Google - eg. poiuytrasdfghjkl 3. Link to both files from the same spot on your website, as to allow them the same chance of being crawled and indexed by the bots. ... *drumroll*... 4. Build backlinks to only 1 file. 5. Search the unique string in Google and see which file ranks higher! TA DA... rinse and repeat to test for consistency. The setup takes far less than 10 minutes, even though I would give the backlinks 2-3 days to index.

Nedim Šabić

06/07/2011 03:46 pm

I see this just as a new way to get information on how to find good linkbuilding methods. If this would really fit in the actual topic, it wouldn´t make restrictions like "everyone can do it". It is easy to prove this. If you get a PR8  link from a realted site, it WILL impact for sure. The clue like I said - of course not everyone can do it.

Michael Martinez

06/07/2011 04:01 pm

I appreciate your post, Barry, but Ron's responses in the forum were trollish and provocative.  He has a long history of being hostile toward me and has on more than one occasion derailed discussions in order to play character assassin.   He made it evident he had no intention of seriously discussing either the issue or the proposal, but rather was only interested in trying to embarrass me. Unfortunately, that pretty much means that all people will find in that discussion is a pissing contest.

Michael Martinez

06/07/2011 04:03 pm

"If you get a PR8  link from a realted site, it WILL impact for sure" Unless the site (or link) has been stripped of its ability to pass anchor text, you're correct.  This is why we need a credible link value test that anyone can use.  Most people don't realize just how little value is being passed by links.

Michael Martinez

06/07/2011 04:04 pm

Sorry.  Unique anchor text is a well-documented link value test. The test I am calling for has to work without anyone creating any content anywhere.  It cannot rely on unique anchor text.

Michael Martinez

06/07/2011 04:06 pm

The reasonable person standard was established in courts of law as a means of differentiating between the two extremes of "only an idiot would believe this" and "you have to be an expert" and the main population, the "reasonable person who may or may not know much about the topic but can bring a common sense approach to it". So there is no conflict here.  A reasonable person has to be able to trust the test's results, and that means 100% accuracy.

Michael Martinez

06/07/2011 04:09 pm

Sam, can you tell in 10 minutes or less whether a Website is indexed, whether it is ranking for any specific keywords, whether the page titles and meta descriptions are unique, whether it is relevant for any specific keywords? I can and so can most other people. So asking for a means of evaluating the ability of a linking page to pass value through its links in 10 minutes or less seems reasonable to me.  SEOs like to talk about how they get their rankings through links but they can't show which links are doing that for them. That's called BULLSHIT.


06/07/2011 04:18 pm

Please, everyone, believe what Michael says so that my clients can dominate your search results.


06/07/2011 04:36 pm

Michael and I used to have a fairly civil relationship online, although we disagreed quite often. However, more recently I have become aware of this and other troll-bait style antics that I feel damage the discourse. That being said, I will respond this time seeing as Search Engine Round Table has decided to take the bait. Michael's rules may be clear, but they stand fundamentally in contradiction to the requisite features of an effective test. Google is essentially a black box. We give data to Google, and Google gives back results. However, because it is a black box, we cannot trace the path of a single variable through that box, unless we can prevent Google from looking at any data other than the single variable. Theoretically, we can do this through controlled experimentation. Even this, though, is suspect as Google probably includes some level of split testing in results, meaning the variation of a single variable must not only be powerful enough to cause ranking changes, but must be profoundly enough such that Google's split tested results do not undo at least part of these changes. Moreover, because it takes time for Google to index new content and adapt search results, a reasonable time frame would be days or weeks, not minutes. Thus, Michael's requirements of 10 minutes, existing links only, and 100% accurate make this an impossible test.   More importantly, Michael presented a false dichotomy from the beginning. He is arguing that because we can't show which links send juice and which do not, then we cannot "believe in the link fairy": that links improve rankings. This is akin to saying that because we do not know which of the yeast cells is creating CO2, then Yeast cannot cause bread to rise. The truth is, it does cause bread to rise, and if we had a microscope - just as if we could look into Google's databases, we could know which yeast cells are performing and which links are passing juice. He continues to make these false relationships such as --- unless you can tell which links pass value, you are wasting your time with Link Research. Just because I cannot tell which yeast cell does the work, I still use it to leaven the bread. Instead, it is a game of RoI. As long as the research and effort provides positive business results, I needn't quibble over minute details of which individual link was valuable and which was not. We can increase efficiency by  attempting to acquire similar position increases over time with less budget. Now this requires some fairly serious multi-variate testing and tracking, but it certainly doesn't require a 10 minute, 100% accurate test. I will leave with this: Please, please please. If you operate in a competitive search engine space, PLEASE BELIEVE MICHAEL. Please, believe that you don't need to work on getting links or doing link research. Please stop all link building and work on your content. Please do this, so my clients can dominate your SERPs.

Michael Martinez

06/07/2011 05:15 pm

The criteria I laid down are quite valid for an inferential test.  You don't need to create any links or content in order to look at the available, existing data and draw some conclusions. It's the SEO community's reluctance to question its own assumptions that have led to this point.  To be honest, up until this year, I HAD a 10-minute inferential test that I was fairly confident in.  All I had to do was run a simple query. I was not 100% confident in the test results, but I was very confident.  Of course, Google changed some things around and the test no longer produces anything that looks like useful information. Testing does NOT require you to create Web pages.  It requires you to look at data.  The creation of links and pages and waiting for indexing all that is just ONE WAY of collecting data (and not a very efficient one at that). These kinds of tests are NOT easy to devise.  I've only used a very small number of them over the years.   But it IS possible to devise them. Whether anyone can do it by June 30, I have my doubts. Whether anyone who CAN do it is willing to share the test publicly, I have even more doubts.  After all, I'm only offering $1,000 but that's all I'm willing to put on the table.  It's enough to show that the offer is serious.

Michael Martinez

06/07/2011 05:18 pm

"More importantly, Michael presented a false dichotomy from the beginning. He is arguing that because we can't show which links send juice and which do not, then we cannot "believe in the link fairy": that links improve rankings. " This required a separate reply.  That is completely wrong and misleading.  I don't say people cannot believe in the link fairy. I say people won't admit that they are waiting for the link fairy. In fact, there is no true accountability for link building campaigns BECAUSE you cannot show which links are passing value. The only false statements in this discussion are those being used to twist and distort what I say.  There is no need to resort to such smarmy tactics in order to disagree with me.

Rank Chef

06/07/2011 08:34 pm

I do not understand this requirement, at all!? 1. I did not say to use a unique anchor text to link the files. It is an identical process, don't use any anchor... just the URL 2. "test I am calling for has to work without anyone creating any content anywhere"...  Is that supposed to make sense? How can you even link to something without creating "content"!?!? Seems to me like Michael did a good job of stirring up some buzz around this "contest" with no actual intention of handing out the prize to anybody.

Rank Chef

06/07/2011 08:44 pm

I whole heartedly agree with rjonesx. The criteria you laid down is NOT valid for an inferential test. You DO need to create links to run a test where LINKING is one of the key variables. This is not to mention that testing DOES require you to create web pages, or content in general. I'm surprised you're unfamiliar to this, but looking at data is called ANALYZING...  You're right - creation of links and pages is simply collecting data, but when you do it to carefully isolated subjects in a controlled environment, you're essentially TESTING. You might as well stop playing dumb, this is the douchiest PR stunt if I've ever seen one... you clearly have no intention of handing out the prize to anybody, and as an excuse will continue to expand on your "laid down criteria" (which contradicts itself mind you)

Rank Chef

06/07/2011 08:52 pm

BULLSHIT suits your little PR stunt better than the "SEOs can't show which links are doing that for them". SEOs can. In fact, anybody who has the time to sit down and get some math done can. It's rather simple, but will take more than 10 minutes, use the attached formula as a guide :) http://upload.wikimedia.org/math/8/0/1/80125f33d12ceb608fdb9daec09d9c10.png

Michael Martinez

06/07/2011 09:15 pm

"You DO need to create links to run a test where LINKING is one of the key variables." In that case, it's NOT an inferential test.

Michael Martinez

06/07/2011 09:18 pm

The point of the test is to enable anyone doing "link building research" or "competitive analysis" to determine whether a linking page can or is passing value. There is no point in obtaining a link that won't pass value (if your objective is to garner more PageRank and/or anchor text). There is no point in including a site in a competitor's backlink profile (when you're trying to figure out how they are ranking for anything) if the site doesn't pass value through its links. You have to understand that most people in the SEO community DO NOT UNDERSTAND WHAT THEY ARE DOING when it comes to links. Did I stir up some buzz?  Yeah that was the point -- not to get more links (that's the dumbest accusation anyone has leveled at me) but to get people to THINK.

Michael Martinez

06/07/2011 09:18 pm

Rank Chef, attacking the people you disagree with doesn't make you right.

Michael Martinez

06/07/2011 09:19 pm

The personal attacks and lies that Russ and others are posting here are completely unnecessary.  If you don't agree with me, just say so.  If you want to argue over what constitutes a valid test, leave the lies and personal recriminations out of it. I'm not interested in the insults and slurs.  I doubt anyone else is, either.

Ben Cook

06/07/2011 11:07 pm

I really wish people would learn to just ignore Michael's self-righteous egomaniacal rantings. He's been screaming the same thing for several years just hoping that someone will listen. He could up his offer to $10M and it wouldn't matter, he'll never be satisfied with any evidence anyone provides and will never pay it out. If you provided a copy of Google's algo, along with signed affidavits from every Google engineer possible he wouldn't believe you or pay you a cent. So please, can we go back to ignoring this nonsense now?

Michael Martinez

06/08/2011 12:43 am

Barry wrote: "..., posted a $1,000 reward for someone in the SEO industry to provide a clear 100% positive proof with evidence that links work to rank sites in Google." Now I understand why the toxic personalities came out of the woodwork.  Links DO help with Google rankings.  The point of asking for the test is to demonstrate once and for all that SEOs who depend on links don't know which links or how many actually work; hence, their analyses and strategies are all flawed. Case in point: At SMX Advanced this week one SEO analytics company announced that JC Penney had acquired 220 EDU links over the past few months.  I found a number of those links were job listings.  I asked Matt Cutts (via Twitter) if the EDU job posting links helped JC Penney (the Twitter chatter agreed that the presentation said the EDU links helped JC Penney "get back into Google's good graces"). Matt replied "no". When it comes to links, the average SEO has no clue which links are helping.  The industry needs to stop being petty about this and admit the truth.


06/08/2011 11:35 am

If Matt had said yes there would have been a surge of spam on EDU jobs. Why do you assume that he is telling the truth?

Michael Martinez

06/08/2011 03:21 pm

Because he has no reason to lie.

Ellie K

06/08/2011 06:30 pm

No, I just scanned through his Disqus history. He is the consummate SEO: Every entry for two years back (I stopped at that point) has that same active URL signature. Many had a second content specific URL too. The foul language isn't good, it is disrespectful to Barry (IMHO). But he (Mr. J. Y.) might be having a bad day, or week or month. Most of his past comments were congenial, no 4-letter words, original and worthwhile reading,  even without context. My two cents re the Link Test Reward: Why did Michael Martinez's offer provoke so much hostility? He's the only one subject to any downside risk. If there isn't a winner, he accrues some credibility, not a lot, and not at anyone else's expense or reputation. What am I failing to notice? The Link Test Reward makes us think, even if only to consider whether or not the criteria are  logically inconsistent, or too generalized.  Contests (well, Michael's offer is more of a "challenge") are accepted practice in mathematics, data analysis and basic research in biology or medicine. And information security/cryptography too, nearly forgot that....!

Rajesh Kumar

06/09/2011 06:05 am

give me a link on usa.gov and I will show you the difference within 10 minutes......

SEO Professionals

08/06/2011 06:46 pm

Perhaps the most important -- and inexpensive -- strategy is to rank high for your preferred keywords on the main search engines in "organic" or "natural" searches (as opposed to paid ads). Search engines send robot "spiders" to index the content of your webpage, so let's begin with steps to prepare your webpages for optimal indexing. The idea here is not to trick the search engines, but to leave them abundant clues as to what your webpage is about. This approach is called "search engine optimization," abbreviated as SEO.


08/29/2011 01:53 pm

Any updates on this?  Sounds like Michael doesn't think there is such thing as link value.

blog comments powered by Disqus