Hacker News new | comments | show | ask | jobs | submit login
How Demand Media Used PR Spin to Have Google Kill Their Competitors (seobook.com)
102 points by tortilla on Feb 26, 2011 | hide | past | web | favorite | 65 comments

> Demand Media further benefited from flagrant spammy guideline violations, like 301 redirecting expired domains into deep eHow pages. People I know who have done similar have seen their sites torched in Google. But eHow is different!

Hearing this is what makes my blood boil as a small startup. Small and new brands are walking on eggshells trying to balance self-promotion without pissing off Google. Meanwhile larger brands regularly practice black- or gray-hat SEO and survive because Google can't as easily remove a larger brand from their results. But if you're small, Google can and will kill you in an instant without recourse. The ramifications of this are bad for the internet -- it means higher barriers to entry for startups.

>Google can't as easily remove a larger brand from their results.

Google was willing to ban BMW for hidden text. http://abcnews.go.com/Technology/story?id=8252259&page=1 Also, the author does not substantiate the claim that ehow benefitted from this tactic.

"Self-promotion" means getting customers to like your site, not tricking search engines into thinking they already do.

Google was willing to ban BMW for hidden text

That lasted, what was it, 48 hours? I have no interest in picking a fight, but for the benefit of other HNers because this really matters for their businesses: there is definitely a point in brand value where the rules change. A one-man operation who pulled a BMW would get terminated without a second thought and would find it very, very difficult to get their old rankings back. Google can't do that to BMW because if BMW can't rank for auto-related queries in Germany then Google looks stupid.

(Public sanction like the BMW thing is more a messaging tactic than anything else. I mean, case in point, we're talking about a single identifiable incident prominent enough to be mentioned in the mass media in a foreign country multiple years ago. Ditto the incident where Google Japan got wrist-slapped for buying reviews. What's the French phrase "to encourage the others"?)

I believe the policy for things like hidden text is to revert as soon as the problem is fixed. (Google attempts to email the webmaster about the problem with whatever email address is available, and also puts a notification in the webmaster console.)

Two addendums on this topic by Jeremy Schoemaker, popular affiliate marketing mastermind, regarding what you discuss - I think he articulates the points well.

http://www.shoemoney.com/2007/10/06/dont-make-google-look-st... http://www.shoemoney.com/2010/04/28/where-my-hatred-of-seo-c...

Well, since you work at Google, I'll keep walking on egg shells... Please don't kill me all powerful Google! :-P

In all seriousness, I liked the Google of 5 years ago that temporarily banned BMW better than the Google of today. I wish that the rules were applied for big and small brands in exactly the same way -- but it doesn't seem they are. When BMW was blacklisted I was impressed -- but this stiff punishment hasn't seemed to happen for JCPenny or Demand Media in the same way it has been applied to smaller businesses.

Another example: these are clearly paid links: http://techcrunch.com/2008/10/17/thank-you-techcrunch-sponso... - if small time bloggers try shit like this they get shut down, but this is TechCrunch, so it slides. I think this shows how there seem to be two sets of rules, one for the established and one for the startup.

Google cares about "paid links" in the sense that they don't want people to use their pagerank for monetary gain. This is clearly not TechCrunch's intent. Similarly, tiny conferences like http://www.alohaonrails.com/ do the same thing, and don't get shut down. I don't see the double standard in that case.

Intent shouldn't factor into Google's algorithm as that's a thin line even Google can't accurately toe without getting it very wrong. Just because these links are a by-product of advertising with a website doesn't make them anything but paid links. I didn't look but those links should be no-followed, and if they aren't you'd be crazy to assume those advertisers are gaining nothing (rank) from them.

Google also buried many of JC Penney's results for linkfarming earlier this month.


It's much easier for Google to ban BMW when they don't run AdSense than it is to shutdown Demand Media and directly impact Google's bottom line.

Google would sooner shut down all of Adsense than compromise its search quality. The former would decrease revenue, the latter would be an existential threat to the company. (Look at Google's quarterly reports if you don't believe me.) Even if you don't believe that Google is trying to do more than make money, it should be sufficient to look at the numbers.

"Revenue" is not a metric used to decide whether ranking changes launch. Period.

Sorry but I don't believe this isn't brought up when considering algorithm changes or improvements. If it's not your department it's someones elses (maybe they are ignored or higher up then you?). You mean nobody stands up and says "Hey this will drop our AdSense revenue by X millions if we do this"? It seems a bit reckless to not consider other impacts of algorithm changes when your company is in fact doing all of this to generate a profit.

It's true. It's never brought up as a matter of principle. In fact, a ranking change that could theoretically reward a site for using Google products would not launch.

Didn't AdWords advertisers get a quality score discount (through higher ad clickthrough rates + free checkout processing) for using Google Checkout?

The official page for Google Checkout to this day states "Google Checkout users click on ads 10% more when the ad displays the Checkout badge, meaning more traffic to your site."

That is a pretty clear-cut example of Google tilting their "algorithm" toward promoting another Google product. And in an area where one of our small clients didn't use Google Checkout, that change simply priced him out of the ad auction. His profit margins after ad costs were roughly 10% & with that 10% cut in relative clickthrough rate he no longer had a place in Google's ad auction. He was forced to use another Google product if he wanted to be profitable with AdWords.

Of course one could say that the organic results are different than the paid results, but the Google editions ebooks ranked quickly in the US after the ebook store was launched, Google only rolled out universal search after they bought YouTube, maps & local results now come with tags that earn incremental revenues from the "organic" search results, etc.

I can't say much about ads. I don't know anything about how it works or what their policies are.

Let's take this statement at face value and see if it passes the 'sniff' test.

Q4 Financials: http://investor.google.com/earnings/2010/Q4_google_earnings....

Lets see two quotes: "GAAP net income in the fourth quarter of 2010 was $2.54 billion"

"Google's partner sites generated revenues, through AdSense programs, of $2.50 billion, or 30% of total revenues, in the fourth quarter of 2010."

If Google 'shut down all of Adsense' because they realized it compromised their search quality, they would have had only $40M in net revenue for the quarter last year. (Doing the simple math of removing AdSense revenue from the picture.)

So you think Sergey and Larry would put up with a 30% drop in revenue to take the high road? We're talking 2.5 billion dollars here, not mouse nuts even to Google.

Shutting down Adsense would be a death blow for Google, compromising search quality is a coin toss. I find it very hard to believe that effects to revenue are not thought about or discussed before making changes, it seems irresponsible and suicidal.

When we discussed this before the algorithm change, people -- including me -- pointed out that the Demand Media relationship earned Google $50M a year, and this was likely to influence their engineering at some level. moultano argued that due to Google's self-organization, it wouldn't work out that way. Lo and behold the change rewards demand media and small startups' blood boil. Just sayin' ...

Hold on... sorry if this is off-topic but I have two older domains that redirect into parts of my new site where the exact same functionality is offered. The redirect is a 301 redirect. I thought this was the right way to do things. Am I being penalized by Google for this? Should I just operate those pages directly on the old domains?

No, that's the right way to do things. The practice he's talking about is buying someone's old domain in an attempt to use its existing backlinks for SEO.

Jason "will do anything for ink" Calacanis recently gave an about face speech claiming people need to step away from the content farm business model, and in doing so admitted that roughly everything he said about Mahalo over the past couple years was a complete lie. Surprise, surprise. The interesting bit is that the start up community - which used to fawn over his huckster PR driven ploys - no longer buys them. Jason claimed to have "pivoted" his business model again, but once again we see more garbage content. His credibility has been spent. And so have his rankings! Sistrix shows that not only is he ranking for fewer keywords, but that the graph has skewed downward to worse average positions.



Looks like more of the same to me...

Everyone likes to vilify the content farms and scrapers (and they deserve it) but the real villain behind all of this is CPC/CPM based advertising.

Can you imagine a world where your attention was sold off based on how long you stayed on a page rather then how often you switched pages? If google wants to fix their search results, they should focus on fixing adsense. The technology to more accurately measure a viewer's exposure to an ad are there, it just needs a trustworthy player to bring it to market. Someone trusted by both users and advertisers.

Google made click/impression-based advertising appealing to both groups and it made them what they are now. It's time to get away from it.

What magical solution do you have in mind that's better than PPC? I mean it's an open bidding market where advertisers pay for specific results. It's pretty damn fantastic for calculating exactly how much each customer cost and I only pay for a result (click).

What's the technology that measures the value of an exposure to an ad?

That is a fairly damning analysis of the farm update.

I think ultimately, sites and pages ought to be ranked based on quality, and not on whether they are a content farm, or even spam and google should be as neutral as possibly in its determination of the algorithm, focusing instead on only ensuring the quality is what prevails, and of course relevance.

What possibly has happened is that google has been forced to act in a rush, under pressure, possibly stress, and so the final product is probably a knee jerk reaction which does not address the underlying problem.

You can not make 100 percent of people 100 percent happy 100 percent of the time. Moreover, if you are such a powerful force as google you ought not to respond to pressure groups, such as a vocal minority, and act in reaction to their specific demands, that only causes problems.

But maybe the question of search has now become so difficult that indeed it has turned into a political self interested game. Youtube is hardly a site I would want to appear in most of searches which are not about videos.

> sites and pages ought to be ranked based on quality,

Unfortunately, that's very subjective - e.g. I consider almost all programming material on MSDN to be of very low quality, yet I expect results there when I search.

Furthermore, assuming an absolute content metric, having several sites with the same "quality" should give the original a high score, and the others essentially zero (unless the original cannot take the load). Algorithmically, you can figure out who the original is by noticing where updates appear first. But that's not a measure of "content quality" - it's a measure of "content farminess".

Can you say oops?

Disclaimer: I work for Google. I do not work in search quality (or on search at all). Below are strictly my personal views.

I can't (and won't) comment on the specifics of this change. In fact, my knowledge of them is pretty much what everyone else's is.

What I will say is that in the last few months there have been several stories about the quality of Google search results in the last few months (eg scraper sites rating higher than the original).

The problem I see with such criticism as leveled with those episodes (and this one too) is that there is the implicit premise that Google's search results and algorithm are static. This most recent change should be evidence of that.

So while Google's search is algorithmic the people who are in charge of it are not. To put it another way: if you try and game Google's system, it will possibly work for a time but at some point, when the problem is viewed as being of sufficient severity to warrant attention, that algorithm will change.

Search, as I see it, is an arms race. SEO, particularly black hat SEO, is on the other side of that. But this isn't as simple as SEO. The world changes over time too. New business models form. New memes come into existence (eg the idea of social search).

So let's assume for a second the OP's argument is sound and that Google has merely killed off Demand Media's competition. If true, there are now a lot less content farms on highly ranked pages than there used to be. Sounds like a win to me. Is it a perfect solution? No. But is it better? Absolutely.

Google's mission is to deliver quality content to it's users. The more people use our properties, the more money we make. We are very focused on the user experience. Gaming our system is, at best, a short term proposition as there are an awful lot of bright and talented people here constantly striving to defeat such attempts.

it took Google how many years to slap down content farms? And they let the biggest offender off the hook...I guess it's just a coincidence that eHow is the biggest content farm and generates quite a bit of money for Google.

technically wikipedia is probably the biggest "content farm".

"Search, as I see it, is an arms race."

Search is not an arms race. If someone provides content of a better quality than others they ought to prevail. So if gaming the system is by producing better quality content, then fair enough.

If, however, gaming the system relies, as it does currently, on simplistic rules, such as keyword density, keyord in H1, and bold, and italic, and within the first few sentences, and all these other rules, then yes it does become an arms race and it proves really that those awfully talented people and bright aren't really much of either.

Rather than getting into an arms race which is damaging for everyone, google would do better to sit back, and ask a simple question, how can intrinsic quality be determined? Links are one way. Another way may be the language level which google has recently announced about websites. On that point, if google is able to determine the level of language a site uses, that is high level, medium or low, then surely google is able to determine the quality of the site?

Maybe, google is just too stuck in one paradigm and cant bring itself back and actually focus on the question that matters, how do you determine quality, rather than make some small tweaks which took the seo industry a grand one day to determine.

"Language level" is not a good metric. This comes up from time to time, with one site or another analysing your "writing level", frequently by re-purposing Flesch–Kincaid readability tests[1]. These tests tell you how hard it is to read something, and the assumption then is that my writing is better than yours because mine is "11th grade" and yours is "8th grade". Of course, having higher entry requirements does not mean my writing is in any way better, and may actually mean the converse - if you had written, for example, an accurate account of quantum physics that was accessible to 8th graders, that would be an impressive writing accomplishment indeed! An example of a site that would be unfairly penalised by measuring language level would be the simple English Wikipedia[2].

[1]: http://en.wikipedia.org/wiki/Flesch%E2%80%93Kincaid_readabil...

[2]: http://simple.wikipedia.org/wiki/Main_Page

Google really should answer to the question why all of these other sites got knocked down but not Demand Media - aren't they the worst offender of all? I'm not sure I agree with the theory that it was their PR that did it. Putting on my tinfoil hat here, but did their recent IPO have anything to do with it? A lot of money is on the line now.

When I saw the Sistrix data I was surprised to see Wisegeek take a big hit, because I usually get some utility out of most of their content (even though it is now plastered with google ads), whereas eHow is almost always useless.

I was not saying that eHow's public relations saved them, but rather it is what caused the backlash against content farms.

Whoops - sorry I misunderstood the nuance

I must be alone here, but some of the articles in eHow were helpful (ie, how to form an LLC in some states). It was very basic, but it was on the money.

Sure, it's a "content farm" but if the advice is good and succinct, it sure beats trying to navigate through a particular State's LLC requirements.

Even a broken clock is right twice a day. eHow's business model doesn't see any difference between worthwhile content and inane babel. Some people do put good stuff on that site, but I don't know why they bother.

I've never found an ehow article of any value whatsoever.

Honest question: Can someone tell me what's wrong with a business model of putting public domain content on the web (like their "Creole cookbook" example)?

Clearly there's a line that has been crossed by republishing existing content - but wouldn't old content be fair game?

I don't plan on doing this BTW - it seems to me the hosting costs would outweigh the incoming ad dollars.

If it's presented as an archival project (like Project Gutenberg), it seems like it's performing a public service. It's when people are less clear about it that it seems spammy: you're republishing a 1903 cookbook, but not being very clear about the fact that that's what you're doing.

Even worse, and not really performing a public service, is when it's republishing old content that someone else has already digitized, like the dozens of sites republishing Project Gutenberg books with no real value-add.

If you are obsessed by creole cooking and have some old books as part of your web content, sure. I cant see them making up more than a small proportion of a great creole cooking site though, you need new content, commentary and so on too. Cook all the recipes in the old books and add pictures and descriptions...

Spent the whole day learning about Oracle functions, queried Google over 100 times, and didn't come across a single BigResource result! Google definitely made an improvement.

Oracle searches for me usually turn on either articles by Don Burleson or the Puget Sound Oracle Users Group site.

Does anybody else feel like shorting Demand Media?

I thought about it... but their stock went up.

HN is a leading indicator.

what is the deal with this noise about content farms? A few years ago Wikipedia was a content farm for me - being very "on-topic" it polluted the search results and the few times i clicked on it i was very disappointed by the content. Fast forward to today - majority of my on-topic search ends with me happily reading Wikipedia. Either i became more stupid or Wikipedia content became much more better. I believe it is the latter :)

Isn't regurgitating content from elsewhere and slapping ads on it, is a content farm? Wikipedia can be a notable exception in that they've never served commercial ads (They had donation ads).

By the way, it seems that content farms are synonymous with blogspam. Examples: https://encrypted.google.com/search?hl=en&q=eHow+was+ori...

>Isn't regurgitating content from elsewhere

how about Wikipedia's "no original research" policy? DemandMedia supposedly pays their authors for writing the pieces - which supposedly means original content. Don't get me wrong. I'm not equaling current Wikipedia and their overall great goal with DemandMedia. I'm just wondering - if the content farm hunt happened several years ago how would Wikipedia survived it.

>and slapping ads on it

well, i'm for one would like it if search engines rated lower any web pages with ads - more ads (like page real estate taken) - more scores taken out. It is like with Congressmen - more money in donations from Shell - less my trust that the Congressmen will support any anti-global warming efforts.

> how about Wikipedia's "no original research" policy? DemandMedia supposedly pays their authors for writing the pieces - which supposedly means original content.

I think you're equivocating over 'original' here. On Wikipedia, it means that everything you write is either directly sourced to another work or basic connective prose and obvious trivial inferences; it's a matter of citations and verification. With DemandMedia it simply means that the prose hasn't appeared elsewhere.

>Isn't regurgitating content from elsewhere and slapping ads on it, is a content farm?

Would that not then make Google the number 1 content farm in the world?

Surely that's not the complaint, here. Holding quality constant, ads will make a site better--for a low-quality site, they'll provide links to people who can actually solve whatever problem the site purports to solve, while for high-quality sites they help pay the bills.

Wikipedia improves over time (or at least attempts to). Because people put time and effort into it. Also, it keeps references and at attempts to be accurate. The aim is to create a wealth of information. Content farms are mostly the opposite. The aim is to make ad money. Making money isn't a sin but who are we kidding, one of these two generates quality content and the other one doesn't.

Maybe Google are about to do a push for Knol.

Nyuck nyuck nyuck.

What kills me is that wikipedia scrapped content out of countless sites on places like geocities which are no longer around — so wikipedia gains the benefits from having the oldest page on said topic.

If you seriously believe that show the evidence. You have the geocities archive and the wikipedia dumps. I think you will not find thats where wikipedia was sourced from. For a start most of it was written much more recently than that.

I have to say, I really loved reading this article. It spent just the right amount of time on each topic, had good links to back things up, and generally gave what seemed like a great overview for someone who isn't well versed in the field.

Good stuff.

Thanks. I try hard :)

Not exactly related directly to article, but interesting thought...

After seeing the headlines of Google decimating companies when it finds they used evil black-hat SEO tactics, I wonder whether one could use evil black-hat SEO on a competitor, artificially increasing the sites rankings for a few months, and have the wrath of a Google ban-hammer upon your competitor o_O

Very evil. I wonder if there are companies doing this currently :P


Sorry if this has been discussed here previously. If so, links please :D

I have no doubt it happens all the time.

It's common practice for "grey hat" SEOs to test on a competitor site first, before using it on their own site.

Yep, this is an oft-used argument for why Google might not aim to penalise sites for having spammy backlinks pointing to it.

I think that the recent high profile cases have been a tad more obvious since a bit of research can find that it was indeed that company (or their SEO company..) who built the links.

I'd imagine that in general, Google would aim to simply discount spammy backlinks for the reasons you mention.

Especially since tools like xRumer can build 100,000s of backlinks per day, and various other ones (ScrapeBox etc) can build 10,000s per day...

I just can't imagine how this can end well for Demand Media when they are an update away from losing a vast majority of their traffic.

I find it interesting that I've upvoted almost every comment in this thread, pro-Google and con. I guess Google search means that much to me. And I would love to see Demand Media punished.

If anyone in Google is interested, you guys might want to check out Jure Leskovec's works on blog influence. I think it has the capability to solve the junk-farming issue.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact