Archives for October 2006

Halloween 2006: Zombie Jeeves

I really enjoy Halloween. Let’s recap. Sometime in 2000 or 2001, I went as our chef Charlie (the one who once cooked for the Grateful Dead):

Matt and Charlie

Then in 2003, I colored my hair and went as punk rock Matt:

Matt Cutts as a Halloween punk rocker

In 2005, I picked up the blade of the six-fingered man to become Inigo Montoya:

Inigo Montoya!

See, it’s not a pirate. It’s Inigo Montoya. I was drawing a blank this year, so I asked for suggestions. One suggestion I liked was Silent Bob. He’s easy to do (just a trenchcoat and baseball cap), plus you get full mobility and full facial recognition. This morning, I tried it out to see what I thought:

Matt as Silent Bob

Not too bad. But earlier this month I was thinking about a halloween costume that was specific to search. We were always told that Jeeves “decided to retire in style and cruise the happy seas.” That’s always been the official story:

Jeeves happily sailing around the world

But what if Jeeves didn’t retire? Someone should Ask™: what if someone decided to retire Jeeves by deadly force? My high-level contacts within the search industry have suggested that JEEVES WAS MURDERED!! I have compiled this infographic diagram to illustrate one potential situation:

Ask(tm), who killed Jeeves?!?

In Halloween memory of Jeeves, I decided to dress as Zombie Jeeves.

Matt Cutts as Zombie Jeeves for Halloween

Rest in peace, Jeeves. Rest in peace. Here’s the typical Jeeves pose:

Zombie in profile

And now spammers can have a zombie after them in addition to Inigo:

Zombie on your trail, spammers!

So, are you doing anything special for Halloween?

Quick bits

MSN/Live lays out their intended marketing campaign for the next few months. (Hat-tip to Barry). They also just added the ability to make your own QR barcodes, which are popular in Japan. Worked great in IE, but in Firefox I only got this and the page never loaded:


Also, there’s a new interview with Shashi Seth about the future direction of Google’s Custom Search Engine.

Firefox 2 out

There’s this really cute puppy that one of my colleagues brings to work sometimes:

Firefox puppy!

Her name is officially Olwen, and someday she’ll grow into her big, big ears:

Firefox puppy!

I refuse to call her Olwen. Can you guess what I call her? That’s right, I call her Firefox. 🙂 Can you see the similarity?

Firefox puppy!

Here, compare with this picture of the original Firefox:


By the way, Firefox 2 is out. It pretty much rocks the house. What, you want more info and fewer puppy pictures? Start reading here. Then download here.

CSE stuff

Eric Enge is quick out of the gate with Custom Search Guide, which is a directory of Google Custom Search Engines.

And the AJAX search API folks whipped up a quick example of blending the Google AJAX Search API and a CSE. It’s got four CSEs on one page (product reviews, price comparisons, forums/message boards, and shopping). Then from the AJAX side, there’s blog, web, and news search, plus a strip of related videos at the bottom. If you haven’t seen the latest examples of the AJAX search API, it’s worth a looksee. My favorite is the mash-up of search + a spreadsheet + mapping on this page. You could also use that page to generate phone leads, for example.

Reading through this discussion, Obli went from “bah, it didn’t work…” to a search engine over US and UK universities in less than an hour. 🙂

Review: Custom Search Engine

Google just announced something that I’m really jazzed about: Google Custom Search Engine. Several people mentioned that Google’s Accessible Search was built by using Google Co-op under the hood. Co-op has opened much of that power up to the public, so that anyone can build a custom search engine.

Most custom search engines (whether it be Google’s free sitesearch or Yahoo! Search Builder) only let you select one site to search, or you can offer websearch. Even Rollyo only lets you search over 25 sites.

This new offering lets you easily add hundreds (thousands?) of urls. You can search over ONLY the sites you choose, or (my favorite) you can apply a boost to the sites you choose, with regular websearch as a backfill. That’s really nice, because if your chosen urls talk about a subject, you’ll often get matches from those urls, but if the user types something completely unrelated, you’ll still get web results back. So it’s a true custom search engine, not just an engine restricted to showing matches from some domains.

You can also choose to exclude results from different sites. As far as I can tell, this happens in pretty close to real-time, even for complex url patterns. For example, I added the pattern “*” and started to get results from the Google directory, so I excluded “*” and the Google directory results went away immediately.

The look and feel is also customizable. Search results can open up on google, or you can do an iframe to open results on your own site. The former option lets you pick some colors and a logo to customize; the latter option lets you fully integrate the search results into the look/feel of your site because you just wrap your preferred chrome around the iframe. If you’re a power webmaster, you’ll want to play with the iframe option.

Lots of other nice features are tucked away under the hood. For example, there’s a bookmarklet (Google Marker) so that if you’re surfing the web and find a site you’d like to add to your search engine, you just click and that site is instantly added to your search engine. And it wouldn’t be based on Google Co-op if you couldn’t choose to allow volunteers to edit your search engine and add new sites if you want. 🙂

When I played with the first version, I wanted to avoid the standard stuff where you plug in 1-2 sites and get a custom search engine that isn’t blood-pounding-ly exciting (“Oh, a search box, and it searches. Great.”). So what I did was take my feeds (I was using Bloglines at the time) and exported it as an OPML file. Running a command like
cat export.opml | grep "title=" | cut -d'"' -f6 | grep -v '^$' | sort | uniq
was enough to get the blog urls that I was reading (not the feed urls), and I threw those urls into the custom search engine.

And just like that, *BOOM* I had a search engine that covered 70+ blogs in the search/SEO industry. If I searched for [bug], it would return search engine bugs, not bugs in general. OPML-import was so much fun that the Co-op folks promised to support it (I know that importing from Bloglines works; importing from Google Reader might still need a tweak to the OPML parsing). It’s nice that every blogger can have a custom search engine that is centered around their interests.

There’s one tidbit that didn’t interest me much, but readers will probably be interested to hear. If you build a custom search engine, the Co-op folks provide a way to share revenue if people click on your search engine’s ads. Offering revshare is a great way to take a useful tool and get even more people interested in it.

Which leads to my personal take: there’s companies that tackle search for specific verticals (Trulia for real estate, Truveo for video, Kosmix for health and other topics, Powerset for natural language, Guruji for India). Those companies work hard to bring something special to their vertical search. But there’s a tier below “I want to get VC money to do vertical search,” and I think this product could enable that. In the same way that AdSense enabled a lot of very good content creation in different niches, Custom Search Engine could help a lot of people who want to make a search engine, but would be happy doing it not-as-a-VC-funded-startup. Anyway, this whole personal take is wild speculation, but it would be neat if it turned out to be true.

I do think that this launch will kick off a lot of opportunity that not everyone will see or understand at first. For example, the first person to make a truly kick-butt search engine about biking will likely start to attract volunteers and traction and first-mover attention, and could very well become the authority search for that niche. I think that this launch could kick off a wave of search over a long tail of niches; rather than a big vertical like “health,” someone could make a search for the much much smaller “health at every size” movement. Or juggling. Or insurance companies. Imagine your favorite niche, which could be as specific as a small-town or as broad as you want.

Give the Google Custom Search Engine a try. It’s easy enough that you can make a search engine quickly, but there’s a lot of power under the hood.