Google’s Custom Search Engines run amuck!
What do these three search engines have in common? They are all examples of Google’s Custom Search Engine (CSE) program.
When I published my very first Top 100 Alternative Search Engines list on Read/WriteWeb on January 29, 2007, comment #22 (perhaps you remember it) was from one Andrew Mitton which warned, “Keep an eye on Google’s custom search engines.”
He was right. A short while back I covered a search engine called Coins2.0. 
Like all the others, all you need to do is assemble a ton of websites related to your favorite topic: coin collecting (4,200 sites!), harness racing, alternative energy, whatever. Then you run it through Google’s CSE program (there are alternative search engines that can help you with that also), paste on an attractive logo, and shazaam! you have a custom search engine. The result won’t have much, if any, spam or irrelevant search results, because you hand picked the sites. Will it be as comprehensive as a good web search? No. Will it suit your purposes? Yes, because you can just keep adding relevant websites until you have created your own complete little index. And as all of those individual sites are updated, your information will be refreshed as well.
What does this mean for the alternative search engines? It adds to the spectrum of Vertical search, which seems to casting a very large shadow on the horizon. On the far left you have the general search engines, probably Google, right? It’s essentially the broadest “Vertical” (yes, I know it is really Horizontal Search) because it encompasses virtually all topics, so just for a second let it be the Meta-Vertical.
Then, moving to the right on our spectrum are the true Vertical search engines: Job Search, Video Search, Health Search, People Search, MP3 Search, Blog Search, and on and on. Just imagine what percentage of the Top 100 Alts are Vertical search engines (some crawl the web, some just search databases - just like a CSE). This month’s Search Engine of the Month, Answers.com is essentially a Vertical information / research engine.
On the very far right of our spectrum are the CSE. Not Video search, I could do ‘Wolf movies search.” Not MP3 search (a division of Audio Search, which is a division of Media Search), I could do “Irish folk music MP3 Search.” Or the Lifeguard Job Search Engine.
Of course Vertical searches, all across the spectrum, are almost always for specific information. For discovery search, you would probably want a discovery search engine, but not necessarily. If I put a Swiki on the Coins2.0 CSE, I could poke around the tag cloud and find things that I did not know about.
What do you think? Is the Search Engine of the Day feature on AltSearchEngines just going to become the Google Custom Search Engine of the Day?
If you mentally spread all that you possibly can along the MACRO to MICRO Vertical Search Spectrum, what’s left?
September 5th, 2007 at 12:15 am
A good vertical search engine isn’t just a search engine that cover a niche of information. A vertical search engine should apply specific considerations (algorithms) when returning results. Searching in news articles or blogs shouldn’t be the same as in discussions. Each resource has different categorizations and properties needed to take into account. Granted, Google searches it all (blogs, news, groups etc.) but it applies the same considerations when extracting information (Page rank) and where page rank works great on web-pages it doesn’t on discussions based sites. It is not enough to divide the information into pieces, you should also handle it differently. So no, I don’t see CSE working in the long run
September 5th, 2007 at 4:50 pm
Google CSE just applies a filter on its index; it does not have any originality and is as good as the author understanding and knowledge about the subject. As a user, I do not trust search results on a CSE and it can not be as good as comprehensive web search. Knowing that Internet today is extremely dynamic, it would be a huge problem for the author to compile a good list of relevant web sites.
I believe Google CSE is not a vertical and should not be categorized as a stand alone search engine.
September 6th, 2007 at 5:46 am
I don’t always agree with you, but I see that you are objective in your
postings. Despite the differences I still enjoy reading your posts and I
often learn even when our viewpoints are different.
September 7th, 2007 at 7:23 pm
I think that there can be no disagreement
that a tool is limited by the skill of the user.
(People have built monuments AND smashed their
own finger with hammers.)
The criticisms I’ve seen on the Google tool tend
to use the a user’s lack of skill as a critique of
the tool itself.
I think that if a site owner is an authority or
subject matter expert with adequate tech skills
as well, she can enhance her users experience in the
following way…
Given that a site has been around awhile offering
content and interacting with a core user group, the owner
should know:
the users demographics, goals and information requirements,
what content they access most,
what sites they arrive from,
where they go when they leave,
the search terms made to get to the site and
the ones they use while there.
With a comprehensive keyword research performed
in addition to what is now known (or inferred)…
I think that owner would have sufficent relevant
queries to run against a meta-engine (Dogpile?)and come up with
a core group of a couple of thousand sites to use as
a foundation for providing the JIT information requirements
of the overwhelming majority of their users.
Users behavior against the core group of sites can be then
used as a basis for expansion.
This, I think, addresses the filtering problem as well
as the base problem of providing a good list of relevant sites.
Admittedly, I only spent 10 minutes working on this so I welcome
any reponses.