Google Dance!

I tend to be pretty wary of large crowds at a party, doubly so of so-called “dance parties”. But in 2003, Google was still a shiny new world to me, and I felt obligated to immerse myself in as much of the culture as I could.

For the SEO – that’s Search Engine Optimization – business, the Google Dance was a time of fear and opportunity. You see, back in the aughts, computation was hard (bear with me kiddos). Computing PageRank, the essential “quality” of a web page that was central to is ranking in Google searches, required doing a massive, massive iterative graph computation on The Entire Worldwide Web. Which, even back then, was big.

Of course, PageRank was only one component in the mystical secret sauce formula that determined ranking (I had to work with the raw code once; that’s why I wear glasses now). But it was a big scary computation, which meant that even with Google’s legions of servers, rankings could only be updated intermittently, on the order of every month or so. And when the rankings were updated, SEOs who had staked their business models on being in the top seven results for “Cheap Airline Tickets” held their breath to see where they’d come out in the shuffle. Some sites went up in the rankings, some when down, and some – if the system thought they were spammy – disappeared altogether.

The Google Dance was a heady, scary time, even for those of us on the inside. You see, for every honest “white hat” SEO who played by the rules of making their pages Google-friendly, there were probably as many “black hat” SEOs who tried exploiting weaknesses in our algorithm to make their sites seem unreasonably relevant or high quality.

Matt Cutts was the head and heart of our web spam detection team, and he had a frighteningly keen eye for spotting these exploits and taking them down. Sweetest, kindest guy you could imagine if you played by the rules. But if you didn’t, he’d make your worst nightmare seems like a day of puppies and sunshine. We’ll get back to Matt in a bit here, but his constant lament was how fighting webspam inherently degraded quality. The web was so large and heterogenous that every time you adjusted the algorithm to take out a blatant spamming technique, it was difficult to avoid affecting at least a couple innocent sites.

It was a tough command to lead, and it was a never-ending battle. You’d discover that some site had found a new technique for keyword stuffing, and add a parameter for detecting and eliminating the cheat. But two months later they’d be back with a subtle twist using passive link farms. It was cat and mouse, with billions (that’s billions with a “b”) of dollars of online commerce – and Google’s reputation – at stake. So the Google Dance was indeed quite a day of reckoning every time it rolled around.

OLYMPUS DIGITAL CAMERA

But “Google Dance” had another meaning as well. The Search Engine Strategies conference used to be held in San Jose each year, and it was there that all the SEOs, white hat and black hat alike, would gather to exchange, well, search engine strategies. Because it was in everyone’s best interests to have good relationships with these SEOs (or at least keep an eye on them), Google sponsored a party – a big dance party – at this conference, and invited folks to the campus to chat, drink, nibble on tasty food, and dance.

As I said, 2003 was the year of my first Google Dance, so I felt some obligation to go and see what the fuss was. There was the expected loud music, jam-packed crowd, fabulous food and brightly-colored props everywhere. But there was also something else: sure, these folks were here to party, but they were also here to try and kiss up to Google engineers and each other under the influence of alcohol to finagle any inside angle they could about Google’s ranking algorithm. It was like swimming with sharks. Smiling, preening, well-groomed sharks in polo shirts and khaki.

I did a pretty good job of only talking with other Googlers, and then only on innocent topics. But somewhere still early in the evening I noticed Matt ambling in the thick of it, cheerful and innocent as if on a walk through the park. Back then he was still mostly unknown to the general public; his one public function was occasionally serving as the anonymous online voice of engineering known only as “the Google Guy.” But the SEOs knew he was somehow critical to their success or failure, and a sort of ripple followed in his wake as he moved through the crowd.

I tried not to get too close – I didn’t want to interfere with his mojo – but I was curious what the interface between Matt and this teeming, writhing industry looked like. I didn’t have to wait long.

I have no idea what the guy’s name was. I remember it being something like Dmitry or Vlad, and remember him speaking in a Slavic or Russian accent. But that’s probably just because most of the Bond villains of my youth spoke in Slavic or Russian accents. And he definitely looked like a Bond villain.

Here’s my recollection of how the conversation went (note: I’m completely BS’ing the technical content):

“Matt! Matt Cutts! It is good to see you, my old friend.”

“Good to see you here too, Vlad. Are you enjoying yourself?” Matt seemed genuinely pleased, as if he were encountering a long-lost comrade, and not one of the banes of his existence.

“Oh boy, am I having a good time. Listen – you think you’re so smart…”

“How’s that?”

“You know how much it cost us to set up that link farm? And we only got what – two months out of it before you shut us down?”

“Ahyup – that one did give us a headache. You guys are good.” Matt was nodding – you could tell he’d been impressed.

“Indeed we are. And you’re never going to figure out what we’ve got going now.”

“Maybe we’ll get it soon enough. What are you doing? Domain variations? We’re onto those already, you know.”

“Ah – but not the way we’re doing them. You only detect variations by Hamming distance. I know; I’ve been probing the algorithm.”

“So you’re doing what? Common misspellings?”

“Even better: phonetic mappings. And we’ve got a shadow registrar, so it even looks like they’re owned by different sites.”

“Nice. Man, Vlad, you’ve put a lot of thought into this. I’ve got to say, you’re one of the best.”

“Damned right, Matt.”

And after Vlad wandered off, beaming with pride in his cleverness, Matt pulled a little notepad out of his pocket and scribbled a note or two on phonetic variations and shadow registrars. Then he slid back into the crowd, cheerful and innocent as before. Perhaps even more cheerful than before: another spam network was about to bite the dust.

OLYMPUS DIGITAL CAMERAOLYMPUS DIGITAL CAMERA

OLYMPUS DIGITAL CAMERA

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s