Do You Hard Code?

Googler Matt Cutts, whose site I recommend if you want to stay current on that which Google allows him to talk about (usually Adwords) has an interesting post on the practice of hard coding a response for a particular query, in this case, "Google." Loren points out Yahoo has…

Googler Matt Cutts, whose site I recommend if you want to stay current on that which Google allows him to talk about (usually Adwords) has an interesting post on the practice of hard coding a response for a particular query, in this case, “Google.”

Loren points out Yahoo has hard-coded some special behavior for the query [google]. Searching for [Google] at the Australia/NZ version of Yahoo returns “Try the new Yahoo! Search:” with a prominent search box right above the listing for Google. Interesting..

This reminds me of similar incidents from other search engines in the past. Remember when Inktomi hand-coded the result for [dumb motherfucker] so that the #1 result was a Google page about its executives? Boy, I do. That was motivating.

But let’s be honest here. Everyone does this. Google might do it through targeted AdWords, but it’s hard coding by another name…

Author: John Battelle

A founder of NewCo (current CEO), sovrn (Chair), Federated Media, Web 2 Summit, The Industry Standard, Wired. Author, investor, board member (Acxiom, Sovrn, NewCo), bike rider, yoga practitioner.

5 thoughts on “Do You Hard Code?”

  1. I remember reading the following article in Wired a little over two years ago. I’ve included the paragraph of interest, and my comments are below.
    [Query coming in to Google] What to tell a suicidal friend
    Stricken, I glance over at Rae, who has returned from night league volleyball, his spiky blond hair still wet. He, too, has seen the query and is typing away furiously. Finally he stops and looks up at me. “They’re going to be OK. They got referred to the right places.”
    “You can do that?”
    “Yeah, well, I can see how the system responds. And if it doesn’t give the right information, I’ll find better sites and attach them for future queries.”
    “But you can’t help the people who ask the original question.”
    “Just the ones that follow?” Rae nods.
    “You’ve just got to do the right thing. The hard part is figuring out what the right thing is.”He thinks a moment, then gestures at the screen. “I know people trust in this thing. They believe it will have the answer. And I don’t want it to fail them.” As Rae talks, 50 more queries scroll up the screen.

    What it sounded like, to me, after reading this article, was that Google was hard-coding a result to this particular query. Now, by itself, this particular instance of hard-coding is not at all “evil” or even that unreasonable. We all would want this person to find the right answer.

    But I showed this to a Google engineer acquaintance of mine, and he swore up and down that Google does not hard code anything. Google might remove pages, but it certainly, he assured me with absolute certainty, would not boost anything in the rankings in a non-algorithmically neutral, hard-coded way. Not even to benefit suicidal people.

    Well, I still didn’t quite believe him, but what could I say? He certainly had much more knowledge than I did.

    A year later, in the fall of 2004, I was still thinking about this. It was around the time that Google Scholar was released. So I decided to try a little experiment. The very next day after Scholar was made public, I went to Google and typed “scholar” as my query. I figured with so many web pages out there, academic sites, journals, and so on, that I would get Oxford University’s home page. Or perhaps the page for Rhodes scholarships or the like.

    What page appeared, ranked first? Now, had I issued that query a month later, I might have expected that page to be ranked first. With all the news and blog article anchor text pointing to this new Google citeseer, it might’ve easily Googlebombed its way up to the top.

    But I did the search the very next day, after the site was announced. Hardly enough time for Google to reindex the web and recompute all those “scholar” links. Hardly time for such a new web site to gain the trust and longevity that helps boost one’s rank. Maybe I’m wrong. Maybe Google’s crawler had priority-indexed these oft-updated news and blogger web sites all the previous night, recomputed the statistics, and given much more trust because itself is a trusted domain.

    But to this very day when I search other terms such as “labs”, I get Google Labs ranked first. And when I search for “desktop”, rather than getting some Apple or Microsoft page, I get Google Desktop Search ranked first. That feels very, very strange.

    Matt Cutts, in his blog, says he doesn’t mind when search engines hard code.. as long as they admit that they’re doing it. It appears, at least to the outsider, that Google hard-codes as well, in terms of self-promotion (desktop, scholar, etc). And perhaps also for victims of suicidal thoughts, as discussed above. But it really does appear that they do it.

    But they swear up and down that they don’t. So, do they or don’t they? And if so, have they ever told us when and why?

  2. I think that anytime a searcher becomes aware that results are hand coded the engine loses credibility, especially so if there’s an obvious axe grinding…

  3. Hmph. Lets try and be accurate here. When Google does it too “through Adwords”, presumably the user can tell that that particular result is not from the organic set of results but has been purchased by some interested party. Deliberately fuzzing that distinction and suggesting that there’s little difference between paid and non-paid content is a slippery slope to a mess that I am surprised you’re not more sensitive to.

  4. Interesting that Matt and others missed the whole context around the [dumb motherfucker] handcoding at Inktomi. It was actually done as a joke in response to Google getting Googlebombed for that term and having the George Bush home page at #1. It was not axe grinding as Andi pontificates- it was a joke and was up for about 30 minutes. Obviously everyone was looking to see if Inktomi was vulnerable to the Googlebomb too at the time when the news broke and so someone put up the google page for this query. Lets not rewrite history here…

  5. Tim, please read Matt Cutt’s post again. It’s not about Inktomi.

    Matt’s post is about Yahoo’s hand coding some results, The Inktomi handcoding was mentioned as an incidental aside, and Matt’s reaction to it was the point of the aside.

    I was making a general statement about hand coding search results. I will admit to pompously looking down my nose at search engines that do this (I try not to be too boring when I pontificate).

    Thank you for adding some detail to the story but I think you should take a deep breath and calm down. It’s not about you.

Leave a Reply