According to Beal, Yahoo’s new CAP (paid inclusion) program is not going over that well at the SES show. I can’t imagine it would – after all, Yahoo is asking folks to pay where before they didn’t feel they had to. The boards are negative on the move, though at least Yahoo is out there explaining itself (and Jeremy adds his two cents here). This might all blow over as the market adjusts to the realities of capitalism, but…
…before Yahoo made this move, I suspect that many marketers could get by on organic Google crawls and targeted AdWords/Overture buys. Now, because Yahoo is in a duopoly position, the stakes are raised, and marketers feel like they *have* to be part of the Yahoo club. They fear, I am sure, that if they do not pay, somehow they will be treated as second class citizens by Yahoo’s Slurp! crawler. Yahoo insists that its organic crawler is going to be “aggressive” and that paid inclusion is an add on of sorts, it insures that your site is crawled thoroughly and in the manner you choose (in particular, you can specify dynamic content, get fresher crawls, get reports, etc.).
But the company seems a bit tone deaf to the more primal forces at work here. First, the timing. Yahoo got a lot of mojo for the new engine it rolled out recently. Why taint that credibility so quickly with this announcement? The company could have waited a month or so, prepared the market for this in stages (for example, it might have cultivated an influencer network prior to the announcement, as I point out in my current 2.0 column). Secondly, it’s predictable that marketers would feel like they have no choice, that they are being forced to enlist in CAP. That is not good for a company’s reputation long term. If I were Yahoo, I’d monitor this closely, and adjust a bit if need be. Perhaps soothing words and assurances will be enough while the market swallows this new dose of medicine. But Google can and will make a PR killing portraying Yahoo as the Big Evil Biased Bully. Ask Jeeves already has: yesterday it cancelled its paid inclusion program …. I am sure the timing was pure coincidence.
As, er, a thought experiment of sorts, take a look at what each search service (Google and Yahoo) bring up when you search for “yahoo “content acquisition program””:
Google
(top two are webmasterworld board posts, questioning the program)
Yahoo
(straight news stories, not nearly as controversial)
Algorithms have no bias? Draw your own conclusions.
Update: Resourceshelf has direct quotes from Ask spokesperson on why they ended the PI program here…
Of course algortihms have bias. Developers make an indirect editorial decision when writing them, so that they show bias in use is not surprising.
But I don’t think a fair conclusion would be that Yahoo’s search results are inferior to Google’s. With a such a general search (yahoo “content acquisition program”) in a general web search engine for current news, I believe that Yahoo did a pretty good job ranking a leading trade journal, a blog and a media site in the top three. It is not the only way to rank, but a pretty good one.
Looking at the first ten results, I believe Yahoo is actually giving somewhat better answers than Google (two directory listings!).
I’m not arguing that one is better than the other, in fact, I think that’s entirely subjective. It’s quite clear that the conspiracist might find a reason to claim that either a/Yahoo is pushing out negative references or b/Google is highlighting them. Of course, the cynic would claim c/ both. I think it’s d/none of the above. But it is interesting nonetheless.
J-
1) Agree with Henrik!
2) Agree with you, from a pr standpoint Yahoo blew it.
3) The press reports make is appear that p-i is something new. It’s not. First introduced by Inktomi(3/7/2001).
See: http://digbig.com/3trx
4) Yes, Google DOES NOT have a p-i program.
However, I would argue that Google results (and those from other engines) are manipulated to some degree by the work of the seo community in figuring out how to maximize placement Google doesn’t directly profit from it but the SEARCHER needs to deal with it.
As Henrik points out, “algortihms have bias” and some people work to figure them out for the benefit of their clients. Doesn’t mean it’s bad, simply the way it is.
5) I’ve read that an entry in DMOZ/ODP helps boost placement in Google (remember that Google looks at many criteria beyond link analysis/PageRank. I’ve heard from a couple of friends that they’ve had issues trying to get their company listed in the directory. Who are the editors who decide what gets placed into DMOZ? How is the decision made if someone qualifies as an editor? What criteria do they use? Do they have a collection development policy?
6) Coming at this from a researcher perspective, problems can often be avoided or reduced if the searcher creates better search strategies and takes advantage of advanced search features. I realize this is not who the engines play to but…
I’m an engineer at a search engine (and I’m not going to say which one). And I have to say that it depresses me when even someone as smart and engaged as John Battelle plays the game of judging engines based on one query.
For any pair of engines, it’s pretty easy to find at least one query that makes one look better than the other. Real tests need many queries before you get any statistical significance.
You know this, I know this, John Battelle knows this. For my part, the only thing I find more tiresome these days than the single-query test is the paranoid variant of the single-query test. E.G. finding a single query related to Linux that MSN search does badly on, and drawing very dark conclusions about bias… This is the kind of example that John gives us, though, in the “content acquisition program” query.
Coyly playing both ends against the middle, John says this:
“It’s quite clear that the conspiracist might find a reason to claim that either a/Yahoo is pushing out negative references or b/Google is highlighting them. Of course, the cynic would claim c/ both. I think it’s d/none of the above. But it is interesting nonetheless.”
Um, John, if you actually don’t believe that this is evidence of bias, then _why_ do you find this interesting, exactly? I mean, I don’t happen to believe in astrology, and as a consequence I find it much less interesting than those who believe in it do. Do you react differently?
TIm –
I understand your response to my post; here’s why I find the results of the one-query search interesting. Yahoo and Google have very different approaches to search results. In a “moment” such as the one at hand (a news story germane to the two companies) many people will be going to one or the other of the engines and putting in the query (or something similar) that I wrote about. (In fact, that’s exactly what I did. I wanted to know more about CAP, went to Yahoo, and didn’t get what I wanted, which was a page explaining the CAP program. So I went to Google, and also didn’t get what I wanted, but I did notice the difference, which is how it ended up in the post).
In any case, if a searcher went to Google, at the top they saw posts which lead one to presume there is a lot of negative response at Webmasterworld. If they went to Yahoo, they saw a bunch of mostly mainstream media coverage. This is a fact, not a theory. That is interesting, because of the perception that is being created by the engine’s algorithms. I believe that search is a medium, it creates contextual reality in the eyes of a person searching. So when one version of reality is negative and another is neutral, well, that’s noteworthy. To say it’s not interesting that Google is highlighting Webmasterworld would be ignoring reality. Google *is* highlighting Webmasterworld. That’s simply true. (BTW, I also think it’s interesting also because it’s consistent with a conspiracist approach to search results. It’s a coincedence, of course. But I’d wager a happy one for the folks at Google. The fact that their own engine seems to be doing their PR work for them is also interesting – observationally, as a writer).
I am not claiming this is intentional. If it were, it’d be deeply evil. I don’t think either company is deeply evil.
And yes, I find astrology very interesting, as it tends to be a substrate around which we can bond that which we don’t understand, or that which we wish for. Search, I’d argue, is beginning to share some similar characteristics in our culture. Again, interesting.
Interesting discussion. A few thoughts:
It
I agree with Chris B that the question of how you would define bias in a websearch engine is an interesting one. Even the most “objective” engine, that tried to return results that reflected the state of the web at any given moment, would be engaging in interpretative bias, just like photographers inevitably do. And, of course, good search engines do quite a bit more than that, if only by imposing a notion of what a good document should be like, and by screening out or demoting “bad” documents like spam.
If this is the question that JohnB is finding interesting, then I agree that it is interesting — especially in Chris B’s formulation that recognizes that it’s really hard to draw conclusions about such bias without getting off your butt and doing some statistical work.
But if what JohnB is finding interesting is not only that in the murky world of people’s perceptions about search engines it happens to be true that people draw paranoid conclusions about search engines based on one query, but also that you can kinda play with that phenomenon if you’re a commentator about search, and induce people to jump to those conclusions by pointing out suspicious examples, while at the same time disclaiming any belief in such a conclusion — eh, there’s interest there for John, I’m sure, but I’m not sure why I should care.
To put it in a clearer way: there are very interesting things to say about how search engines actually work, and also there are interesting things to say about how people perceive them. I’m just not sure that things get any more interesting if you deliberately conflate the two, and try to create ambiguity about which thing you are commenting on. (And before I put any money down for your book, John, is this example a fair preview of the kind of interest we can expect from it?)
This is a blog, not a book, and as such it’s a conversation. I certainly find the first question (how do you define bias in a search engine) interesting. But I’m not being clear on the more cultural point here, and certainly my initial post was not clear on what I meant when I said “draw your own conclusions.” I was not judging search engines. I was pointing out the power of engines to create cultural context in one person, at one time, with one query. I should have used less inflammatory language, but I was assuming my readers knew the context in which I was writing. I apologize, and realize I should not make that assumption, and I’ll try not to in the future.
But at the risk of boring you, here’s what I was thinking. For most typical users, the very idea of “drawing paranoid conclusions” is not the point. I’d argue that most people don’t think about inherent bias when they are doing a search. As with most media habits, I’d wager that most readers or viewers don’t spend a lot of time decoding why an engine returns the results it returns.
I find it interesting that when reader hears about something he/she want to know more about (in this case about CAP), often the first thing that reader is go to an engine, probably the one he/she uses all the time, and punches in that term. The SERPs that reader sees are the first, instant contextualization of that term for the reader (and I’d argue that often, it’s the only contextulization the reader bothers to obtain). I don’t think most searchers think consciously about the engine’s bias, nor do they compare results from a lot of different engines. They aren’t drawing conclusions about that engine. They are getting a quick shot of context, choosing one or two URLs to click on for further info, if they feel they need to, and moving on.
The point is the power search has to create cultural context. It’s why, more than two years ago, I told Eric that I thought Google was a media company (at that point, he disagreed…).
As to the specific example, which I still think is very interesting, I in no way am drawing larger *conclusions* about any inherent bias Yahoo or Google might have – you’re right, that can’t be done absent a far more rigorous approach. What I was commenting on is that it’s interesting in the sense that a reader can and probably would draw completely diffferent conclusions about *the CAP program* based on whether that reader uses Yahoo, or whether that reader uses Google. The point is not whether the engine is biased, the point is how subjective media is, period.
In my example, I am imagining a typical reader, on the day in question (as results change over time, we need to “freeze time” for this example), who just wants to know what’s up with CAP. If that reader uses Google, he or she will immediately sense something is awry – there are two posts from WebmasterWorld which indicate controversy, and they are the #1 and #2 listings, which means Google “thinks” they are the most relevant. With Yahoo, the reader would not get that initial sense, but would see news reports, first and foremost. Now, the news reports certainly report on the controversial elements of the program, but the initial hit is different. (To my mind, as a searcher, it’s impossible to defend whether the first two links on CAP should be Webmasterworld. If the engine were perfect *for me in this example*, the first link would have been Yahoo’s page explaining the program. But that’s just *for me*….) I think this demonstrates how engines can create context, and also how subjective this all is, which then points to the Holy Grail of personalized search…etc. etc. etc.
Again, I don’t think the reader is (or even should be) drawing conclusions about the *search engine* while digesting the first five or so posts returned back (which honestly is all most readers do). The point is the example shows that engines creates context for the reading public, which is a power that until recently was reserved for the media. I think that’s really interesting.
As for my book, it will be a result of exactly these kinds of conversations. That’s why I created this blog. I hope we can continue this one, and I appreciate your taking to time to force me to explain myself. If you work at one of the engines and have time to meet, I’m down there a lot and would love to talk.
I think Yahoo! using paid search is more of an admission of defeat at math. Hands down Google has built a brand based on math. If I were Yahoo! I do not think I would try to overpower Google fighting them at their strengths.
Thus Yahoo! is using various paid editorial services to give the world a different type of search – a hybrid. As stated above, neither is right or wrong…they are just different.
Paid Inclusion is just one way Yahoo! is paying for the large editorial work at hand. After all, Yahoo! originally was a directory 🙂
cool conversation by the way…