Blog Archives

Blekko Makes Relevancy Improvements, Adds Bunches of Slashtags

Blekko's Monte Tag Shows You Three Different Search Engine Result Options

Blekko's Monte Tag Shows You Three Different Search Engine Result Options

I first covered Blekko about a year ago (November 2010). After that aside from mentioning it a few times I haven’t talked about it, but I still like it consider it to have a lot of good functionality. Recently Blekko announced improved search relevancy and a lots of automatic slashtagging.

Automatic application of slashtags (a way of categorizing information) has been applied to over 500 categories. Using slashtags, you can narrow down your search results contextually, which will get you a wider variety of more relevant results (ideally) than trying to narrow down using specific keywords.

Some of Blekko’s slashtags act as special syntax; using the slashtag /monte at the end of any search query will show you search results from three different search engines. The twist that you don’t get to see which search engine produced which results until you choose the column of results most relevant to you. This is an interesting tool to use because you’ll see a) which results are tagged with which Blekko categories, and b) which sites on other search engines are banned on Blekko. (First impression: Blekko does not like eHow nohow.)

(I also like Monte because it emphasizes, to me, that search has not been solved. When playing with /monte I did not consistently pick one search engine for the best set of results. I didn’t even pick one most of the time. There’s still work to be done to make Web search the best it could possibly be, and nobody’s figured it out yet. Not even Google.)

Blekko Will Make Slashtag Suggestions

Blekko Will Make Slashtag Suggestions

Running a few slashtag searches did show that it was useful at narrowing down results, but the downside is that I’m apparently not very good at guessing slashtags. There is a partial directory available at http://blekko.com/tag/show#tab3. If you want more suggestions you can also enter your search and a forward slash, and Blekko will suggest a variety of slashtags for you.

Blekko continues to impress. The only complaint I have is that I can’t find a complete slashtag directory, either for categories or for special syntax. Perhaps I’m looking in the wrong place?

Blekko Leverages Like; I Embrace My Crabbiness

Blekko (available at http://blekko.com/) announced last week that it has added the ability to use Facebook “Likes” to leverage searching its Web index. Let me explain how that works and give you a couple of examples, and then let me complain a bit.

Blekko’s search enhancement consists of “slashtags” that allow you to sort and/or delineate your searches in a certain way. In the case of this new offering you can connect to Blekko via Facebook, do a search that includes the slashtag /likes, and get as your search result only pages which your friends have “Liked” on Facebook.

(Names of people who “Liked” content blurred for privacy.)

Search results look the same as they usually do on Facebook, except for a small note showing which of your friends “Liked” particular content. Also at the top of the search results you have the usual options to sort results by date and relevance, but there’s also a little “Thumbs up” icon that lets you sort results by how many likes they have received by your friends.

In doing some experiments with this, I was surprised how many Web pages my friends have “liked.” It appears that if a domain is liked, all pages are liked — so if a friend “Likes” the New York Times, all pages on the New York Times are considered “Liked”. It would be nice if you could turn that off, though it would seriously, seriously cut down your search results.

I like this as an addition to Blekko’s search and it seems like a simple way to cut down your search to low-spam, high-quality results. My concern is the proliferation of buttons that allow you to simply like (have you noticed them on Amazon?). The whole spectrum of human emotion against the information and culture of thousands of years, presented by the Internet, and our reactions are to LIKE or not?

Feh.

There’s much to be said for simplicity but I can imagine a “Dislike” button helping a Web search as much as a Like button. Filtering out pages that your friends have specific warnings or concerns against would help a search too, wouldn’t it? Of course you could keep going with this idea — with a “Meh” button, a “Scholarship” button, a “Lolz” button, etc etc. Maybe that’s too much. But wouldn’t be nice to at least express some kind of negative, no matter how mild?

Blekko Joins the Search Engine Wars

I’ll have a review for you momentarily, but I just wanted to announce off the bat that I’ll be suing Blekko for copyright infringement.
I’m pretty sure I have ample evidence for prior ownership of Blekko as intellectual property; the word was used repeatedly in reference to my personal “brand” the last time I tried to serve vegetarian hamburgers at a family dinner.

Just kidding, of course. And I guess there’s some kind of requirement that search engines have odd names. At least Blekko is easy to spell. Blekko has actually been rumbling around for several months now, but launched its public beta yesterday at http://blekko.com. It’s gotten tons of coverage; maybe the search engine wars are revving up into their third cycle. After spending a little time with it, I gotta say: I like. I like very much. I’m a little worried about some of the search algorithms, and there’s a LOT going on in that front page.

But Blekko is letting me get under the hood of search results in a way that I’ve never seen before. I’m getting crazy amounts of data just from a search results page. I’m going to have to think about it and digest it, but the transparency of the SEO information available feels immediately like a reproach to Google. And the slashtag method feels like a reproach to DMOZ and any other searchable subject index that’s still sticking it out nowadays. The whole engine is like a hybrid of crowdsourcing, searchable subject indexes, and the ocean of data more appropriate to a full-text search engine (Blekko’s first search index is the product of a 3-billion page crawl.) There’s some hint of a clustering search engine in there too…

Blekko has the obligatory Minimalist Search Engine look; a search box and some discussion about slashtags.

Slashtags are Blekko’s big thing. They’re both a way to focus search results and a search syntax. If you have a Blekko account (they’re free) you can create your own slashtags and slash the Web your own way.

Blekko has launched with several hundred slashtags and the best way to get a sense of how they work is to try them. Start your search with a regular keyword — say, diabetes.

Your search results will reflect pretty standard Web sites for diabetes. But you’ll also see a number of recommended slashtags relevant to your search. If you want more information about diabetes as it relates to your diet, you can try a search for diabetes /diet (slashtags are always prefaced with a /, natch.) This will rerun your search for the diabetes keyword against a list of Web sites that have been slashtagged for “diet”. And you’ll end up getting pretty relevant results.

It looks like you can invoke multiple slashtags in a search, but it’s going to have the effect of searching those sites which appear listed under both tags. For example, the search for diabetes /diet /nutrition finds you results — but only 27 of ‘em.

If you search for a slashtag that doesn’t exist, Blekko will a) recommend some alternate ones and b) give you the option to create your own. If you search for a tag that someone else has created, you’ll have option to search their tags. (Try diabetes /australia to see an example of that.

There are special slashtags that let you narrow and sort your results in a certain way, but let’s put those aside for the moment, becasue I want to address the search results. Whereas many search engines have gone more and more toward giving you a minimum amount of data for each search result, Blekko busts the results wide open, providing huge amounts of information for each result. Let’s take a close-up look at a search result.

There are several information links for each search result. Taking them one by one, here’s what they look like:

Tag — Add a slashtag (or multiple slashtags) to this search result.

SEO — Get CRAZY, CRAZY amounts of SEO information for this search result page. You will be able to see charts for inbound link distribution, crawl stats, the number of inbound links, and the number of site pages. You can compare two domains and see what other sites are duplicating this site’s content. SEO is not really my forte; if you want a more thorough overview of the SEO tools, you can check out this article at Search Engine Land.

Links — Inbound links for a given result page.

Cache — Shows a cache of the page with your search terms highlighted.

IP — Allows you to search for all pages/sites hosted on the IP of the result page.

Chatter — Make a comment about that result page.

Spam — This isn’t an information link but an action link. If you click the spam link, you’ll remove the page from your search result and you will never, never see it again.

If you click on the prefs link on the upper right part of the page, you’ll have option to see other information in your search results, including Source (shows the page source in browser), RSS (shows a link to the RSS feed, if any), and Similar (shows sites similar to the one you’re looking at.) There are also occasional unusual information links, like Adsense, which lets you search for an Adsense code across multiple sites.

This is so thorough that I missed some things I would expect. I don’t see page size here. No thumbnails. And there’s no JavaScript-powered, in-results page preview. Why not? There’s so much information here, why not just a little bit more?

Okay, let’s move back to those slashtags. Some slashtags (like /diet and /nutrition) allow you to narrow the focus of your content, but others allow you to change how the results display. Appending /date to your search will allow you to sort your results by the most recent additions to Blekko. (diabetes /nutrition /date).
/images and /videos limit results to images and videos, respectively. I didn’t have much success combining these slashes — searching for diabetes /images /date didn’t seem to show results in order any differently from diabetes /images. Appending /rss to your search gives you the results as an RSS feed.

Rich Skrenta has taken the idea of search engines and given it a twist — several twists, actually. I am trying to figure out how to add Blekko to my daily routine, to slashtag resources as I review them and add them to morning buzz. I realized while playing with Blekko that it’ll be an excellent bookmark manager for me. (Don’t tell Blekko; I don’t know if they intended that.)

But I’m a bit worried about the essentials. Sometimes the search results I got we just weird. Searching for strawberry gave me “Strawberry Perl for Windows” as the first result. Searching for “Cow” gave me the Center on Wisconsin Strategy as the third result and a vegetarian restaurant guide as the fourth result (okay, that one was pretty funny.) So the algo needs a little bit of tweaking. (Looking at Google’s results I see that those two search terms also have their own strangeness. I need to do some more experimenting.) But if enough people get on board with adding slashtags and eliminating spam, this could be a very clean, well-organized index.

I would also (I can’t believe I’m typing this) like to see more social aspects to the search tools. You can follow individual users’ slashtag sets, but not all of a user’s content (as far as I can tell), you can’t upvote or “like” user-created slashtag sets (as far as I can tell.)

A great, great start. I can’t wait to get more time to play with Blekko.

Follow

Get every new post delivered to your Inbox.

Join 3,846 other followers