Shanti’s Dispatches

This article comes a little late, after the brouhaha has died down for a few days.

If you don’t know what I’m referring to already, on Thursday Ben Hammersley posted an article, basically knocking the notion that we should be remotely excited about Google’s comment spam / “nofollow” announcement.

Here’s Ben on the announcement and subsequent blogosphere reaction:

Iím deeply mystified by the hallelujahs bursting forth about Googleís rel=”nofollow” method of preventing comment spam. The idea being that comment spammers will leave your own site alone, or stop spamming altogether, if they find the rel=”nofollow” tag. (emphasis added)

Wrong, Ben. Wrong, wrong, wrong.

This is really important stuff here (seriously), so let’s be careful not to mince words. We’re talking about nothing less than how we come to find and access information. How we impart relevance to other websites. How we let the world know when we’re linking to something because we’ve deemed it a valuable and worthwhile resource, or because some greasy spammer in his basement is hammering your weblog with comment spam.

You now have the choice of whether or not to impart PageRank, the fundamental currency of web relevance, to other websites.

You have that choice when you are linking to sites. For example, I’ve actually included a rel=”nofollow” when linking to Ben’s article, because I don’t want to artificially increase its PageRank. You can now feel free to link to competitors, scam sites, spammers’ sites, etc. without fear that you will be passing along PageRank, and boosting that site in the rankings.

If you think I’m overreacting, welcome to the world of SEO. I agree, this has all been blown out of proportion. But it’s the marketplace. In case you’re wondering what the current rate is for a link, go ask these guys. Three measly links (from sites with PageRanks (PR) of 5), for $100 per month. At least that’s what they were charging a few months ago. Ridiculous? Yes. But that’s the kind of math / economics we’re dealing with now.

Now that I’ve explained a few of the merits of the “nofollow” option, let’s continue debunking Ben’s article. He continues:

I think this is false assumption. Itís based on the idea that producing a link costs something, and therefore the spammer must choose either way. Think of it this way: if Iím a guy trying to pitch a TexasHoldíem site, my aim is to get people to go there. Whether this is directly, through a click through, or indirectly via Google, the effect is the same. As a spammer, I donít care at all *how* they get to my site. I just want the eyeballs. The same for any of the other comment spam subjects. PageRank isnít an end in itself, itís just the means by which they get more readers indirectly.

The false assumption he’s referring to is that comment spammers will stop, after hearing this announcement.

Ben - do you honestly believe this is the assumption (from the people issuing “hallelujahs”)?

I don’t know anyone in their right mind who believes the current comment spammers will cease their operations after hearing this announcement.

Will comment spam be taken to the level of email spam — will the meme continue to spread rampantly that you can get rich simply by spamming millions of blogs — will we see one-hundred times the current level of comment spam in two years??

That’s where we were headed, had Google done nothing. Will this put a dent in the profits of the comment spammers, if implemented in a widespread fashion? You bet your ass it will. By how much, that’s anyone’s guess.

It’s easy to say, “Oh, well, spammers will still spam because people will click through on the links, even though they won’t be getting any PageRank out of it.” Sure, some will. But from my experience in the SEO world, these guys are getting 90%+ of their value from the link itself, and not the actual click-thrus. You might think of it as a lottery / gambling system, that they have going.

According to the Overture keyword selector tool, 162,582 people searched for “texas holdem” in December 2004. Sure, a few random stragglers will stumble upon blog posts and click through on your “texas holdem” links at the bottom.

But the real jackpot comes when you get ranked in the top 10 for “texas holdem.” What’s the payoff? 162,582 visitors (Yahoo + Overture’s partners alone) choosing between your link and nine others on the first page of results. I don’t know the exact breakdowns on how many click-thrus you’ll recieve, but it’s fairly obvious that the real intent (90%+ of the motivation) is PageRank, and getting ranked in the top 10 for your keywords, and not the random straggler clickers.

Will this effort actually work? Who knows. But Google should have been on this two years ago. It’s may be too little, too late. But even if it reduces the spammers’ profitability by as little as 20% - then hey, I’m all for it.

Let’s continue. Ben says:

Spamming costs nothing, so spamming sites with rel=”nofollow” doesnít bother me.

Wrong again. Simple Economics 101. Sure, after fixed costs (like Internet lines, proxy server costs, servers, etc.), all it costs is time. And even that is probably minimized by comment spamming robots that do all the work. But these guys’ fixed costs are far from nothing.

First, I can guarantee you that the heavy comment spammers are running fixed costs north of thousands of dollars per month. How do I know this? They are spamming sites from hundreds of different IP addresses, with disperse address ranges. They can only be doing this using a distributed network of proxy servers, or something along those lines. If you’ve ever looked into this (don’t ask me how I know this) - it’s expensive.

To be doing the kind of heavy-hitting these guys are doing, it could be as much as $10k per month. Probably less, but a few thousand dollars is totally realistic. Far from nothing. Maybe they’re using hacked zombie PCs. But either way, this network isn’t cheap. (Even using hacked zombie PCs has its costs — paying off hackers, trips to the psych doc for more anxiety pills, etc.)

Again - I’m only talking about the big boys. I’m talking about the top 20% of comment spammers that account for 80% of the comment spam. This figure is probably more like the top 5% of spammers account for 95% of the spam, but either way - it definitely does not costing nothing - that’s just a fucking crazy assumption. And it’s wrong.

At this point — let’s be clear — I think we’re really talking about keeping even more spammers from entering the game. Sure, the current guys have their fixed costs. They’ve invested heavily in comment spam bringing them some ROI.

We’re really trying to prevent Cletus, that out-of-work uncle you know, who’s sitting in his motor home down in Florida, from jumping on bandwagon because he heard through the grapevine how easy it is to get rich by spamming weblogs.

Later in the article, Ben addresses Scoble’s points that I mentioned earlier:

Technorati will have to choose if itís a site that measures raw interconnectivity, or some curious High School metric of look-at-that-person-but-donít-pay-her-any-attention that the selective use of the rel=”nofollow” attribute will produce. For many purposes, this would mean the results are totally debased and close to useless.

Huh? Wtf? “Totally debased and close to useless” - because you now have an additional metric of information to add to the equation? I just don’t get that. First off, it’s just common sense.

But if you want to get into some kind of crazy Information Theory gambit, then you could go down that road. Any time you have additional information on which to base a decision, you are better off. If you’re talking about a User Interface issue (i.e. how to explain to Grandma that this link isn’t as relevant because it contained a nofollow reference), then yeah, I see your point.

You can bookmark this, digg this or check the cosmos