May 25, 2008

Making My Blog Search Legit With Lijit

Blog widgets are seemingly a dime a dozen these days, but offering a strong search function on your Web site is a must, regardless of how cleanly you've laid out your archive pages, or how well you've implemented tags or labels. A little over a week ago, after seeing Lijit growing its presence on many other blogs I follow, I integrated the service into my site, letting users find older stories I've written on topics they find interesting, and opening up yet another box of stats for me to play with, including most frequently used search terms.

Looking back in my e-mail archives, it looks like I first signed up for Lijit back in June of 2007. If I remember correctly, I think I implemented it, but later, it got pulled in some blog redesign. This time, it's likely here to stay. On May 16, I undoubtedly polluted the Lijit user database, signing up again and getting a second account. Oops.

Acting as a front-end for Google Blog Search, Lijit places a simple search box on your blog, letting visitors search your archives, but also, it pre-populates if somebody arrives on your site from having completed a search elsewhere. For example, if I do a Google search for "FriendFeed Tips", and click on FriendFeed Friday Tips #1: Five Ways To Use the Hide Function, Lijit helpfully asks, "Looking for more about friendfeed tips?" and gives what it would provide as the best links in my blog, as well as through content delivered from other services I use around the Web, such as MyBlogLog and

Search results from Lijit are displayed as a pop-in window in the Web browser, not asking you to leave the site, but instead, showing you the results, surrounded by Google AdSense. Of interest, Duncan Riley said on FriendFeed yesterday that not sharing the revenue with the bloggers themselves was "not cool", but I hadn't given that much thought before implementing.

The most visible benefit of using Lijit is showing site visitors what the most popular searches are, either on my blog, or used to find the blog. As of today, the top ten terms are: FriendFeed, Twitter, Blogger, Lijit, Techmeme, BlogRize, MyBlogLog, ReadBurner, FriendFeed to Watch and Duncan Riley.

Also very helpful is the ability to filter what is displayed. I've mentioned before that there's a core element of Web perverts who like some pages in my archives, so I get all sorts of odd traffic from dirty keywords, which I don't want shown, so I can hop into my Lijit page and add these unwanted terms to the filter.

Meanwhile, as Lijit is watching my site traffic for search activity, it's also monitoring standard blog tracking tools, including page views, how many visitors are coming, and where they are coming from. Combining the two facets of the service, from search to statistics, Lijit can tell me which countries search for what most frequently, what is the city that offers me the most visits (Mountain View, CA), and from what country did my most recent search terms originate. I can also see which pages proved most popular after search terms were entered.

So it works. Good stuff. And while I underplayed Lijit's integration with other services like MyBlogLog,, Flickr, Reddit, StumbleUpon, Disqus, Digg, YouTube, etc., this element may become more important in the future, as site visitors can do more than search just my blog, but they can search all across my network, essentially acting like FriendFeed in reverse, not looking for one site to track my activity, but instead a search point to analyze all my activity around the Web. I'll be watching this to grow over time, and hope to report back and say if site visitors are doing more than searching my blog, but searching my content as well. I've enabled a dozen different sites to pull from, so have at it.