If you haven’t heard about this, it’s a cool idea. Scoble has created a seedbed for testing speed and comprehensiveness in blog search tools, and along the way, has found out other interesting things about search in general. From the post:
Damn, brrrreeeport is the top search on Technorati and there are 420 posts there. Wacky.
What’s an even better deal is that Google says there are now about 14,000 results. What the f___? I HATE the lies that are going on on search engines. Quick: click through and tell me how many entries there really are. Hint: it isn’t 14,000. Funny that Google’s blog search can only find 382.
MSN says there are 1,369 results. Yahoo says there are 1,010 results.
6 thoughts on “Scoble’s brreeeport test”
While what I am about say isn’t completely related to the this “14,000 results” lie, as a user I’m just as frustrated at another type of Google antipathy toward its users.
My experience deals with the appearance of the “In collaboration with NBC Olympics” video that appears when one searches for “Olympics”. This was mentioned a few days ago on this blog.
Well, I did this search again today, and I am seeing not only the sponsored links on the right, but I’m also seeing a blue-boxed sponsored link above the organic results for Chevrolet. So it is clear, I think, that the olympic video box is not “sponsored”, per say. Otherwise, it would also say that, like the Chevy ad.
What it feels like, to me, is that the box is a vertical search result for the video category. Whether video deserves to be ranked above text is a matter of debate, but for the moment, assuming that this is not actually an non-demarcated video advertisement (that would be evil!), the only remaining conclusion is that it is the vertical results of a video search.
Now, to get to my point: I modified my original Google search. I no longer searched for “olympic” by itself. Instead, I tried the following searches:
And in the results to every single one of these searches, the videos still appear. Even though the text of the box in which the video is contained has every one of these words, and should therefore be filtered out or removed from my search results. I specified that I did not want results, text or otherwise, that contained the terms “NBC” or “Torino” etc.
So as a vertical search engine, Google is completely ignoring my request to filter out results that I do not want. It persists in showing me the video.
This is unintentional ineptness at best…some programmer forgot to integrate the “-” operator into the video engine, or it is intentional evil at worst: Overriding my search request to show me something that they want to sell me on, anyway. Either way, as a user it makes me very, very unhappy.
So whether Google is lying about the 14,000 results, as Scoble mentions, or they are (intentionally or unintentionally) ignoring my -video, -NBC, -Torino search request, I end up feeling pretty jerked around.
One last irony? I tried the query “Olympics -chevrolet”. It still shows me the Chevy ad at the top, but the olympic video box is now gone. WTF?
I watched a Channel9 video a while ago where Scoble interviewed some of the MSN search team. In it I thought they mentioned that the returned results number refered to the competitiveness of the phrase rather than how many results there actually were.
Perhaps in this case Google thinks this phrase is relatively ‘competitive’ (whatever G means by this…), hence the high number? I’ve noticed that Google normally returns much higher matches than MSN and Y for the majority of searches, so I tend to use the number returned as an indicator than an ‘actual’ figure.
Its been interesting watching it spread around the blog world and beyond. We wrote about the search results stats issue here
jepp, that is the problem with google’s site seek.
nobody point why google all sides does not indicate 🙁
but thanks for the link 🙂
This brrreeeport test has been very interesting. However, I think it’s now run its course. Rather than tracking the path of brrreeeport sticks as they float down the various search engine streams, people are now engaging in tagjacking. Perhaps there’s a better word than “tagjacking” to describe this notion of hijacking tags. I’ve seen numerous blog posts tagged with brrreeeport and krugle and other top tags that have nothing to do with those tags. Clearly, these bloggers are just trying to get a few more visitors to read their blogs.
This is a potentially serious problem for tags. Will they go the way of meta keywords tags on the search engines? Or, because the tags are visible to the end user, will it be more apparent who is tagspamming or tagjacking or whatever you want to call it and will these people then be ignored?
It´s interesting to see how google get lower resolutions than the page himself, but I think it must be normal. So if you see more then 10000 results it coeld be a well done site or the content must be interesting. Decide by yourself is the better way.