With the CSE, Google allows publishers to specify or prioritize sites to include in searches which are initiated from a search box on publishers' sites. If the publisher wants, other users can contribute to defining its CSE's search results.
I'm not convinced CSEs will be one of the more widely used Google offerings. Still, there are some benefits, and Google's not the only one to think so; others have similar services. Below, I've enumerated some pros and cons of Google's CSE, and I'll leave it to you to decide if it's right for you, or if you think CSEs have legs.
Pro: The G-Lo Effect. There's a formula that keeps reasserting itself: add Google's brand name to anything else, and anything else looks better for it. It's a new twist on the classic halo effect. Add Google to the transaction process, and Google Checkout increases conversion rates. Add Google to a 3D atlas, and suddenly there are millions more budding cartographers. Add a Google-branded custom search bar to your site, and the use of your site's search box keeps growing. That's one of the promises, at least.
Con: A Google search box in the wrong place can serve as a distraction. For instance, a retailer would not want a Google CSE on its main page, since the goal is for consumers to make a transaction on their site, not treat the retailer as a portal. (As an aside, that's why I'm eager to dive deeper into Amazon.com's launch of the pay-per-click advertising service ClickRiver, which will serve ads on Amazon. It sounds like it would distract consumers from their main objective.) There is one way retailers can use CSEs: if a retailer has a microsite, blog, or other online touchpoint that serves as an informational resource for consumers, a CSE could be a welcome addition.
Pro: The CSE is the new and improved version of vertical search. A publisher that uses one is basically developing its own vertical search engine. If I wanted to create a celebrity gossip search engine, for example, I might throw in People.com, US Weekly, TMZ, Gawker, Perez Hilton, Egotastic, and a few others. Someone searching for the latest dirt on the stars would have all their favorite sources in one place, with none of the irrelevant sites. For instance, a Google search on Tom Cruise brings up fan sites, movie pages, photo galleries, and other sites that don't matter to someone who just wants the juicy dirt. A CSE can cut the clutter.
Con: Paid search targeting could suffer. The focus of the CSE results is really on the natural results, not the paid. That's because the natural results become more relevant, while nothing changes with the advertising. For instance, a search for "laptop" in the CSE on Macworld.com brings up search results predominantly related to Macs and iBooks, while the results in Google are all over the map. Yet the ads displayed for the Macworld CSE search are the same as the top ads from the Google search. Given that this includes advertisers only or largely selling PCs, the ads don't provide the same value and even take away from the value of the CSE search.
Pro: CSEs validate an existing model-like Rollyo, a site John Battelle has been touting since its inception, which has offered customized search engines for some time now. There's also Eurekster's Swicki and, the Yahoo Search Builder, which launched in August.
Con: I've had a Rollyo search box on my blog for months, though I'm not sure anyone has ever used it (myself included). As for Yahoo! Search Builder, it made some noise when it first came out, though I didn't hear much about it until Google launched its CSE site. Blogpulse confirms posts about Yahoo Search Builder spiked with Google's CSE announcement, but Google's announcement received 10 times the peak for Yahoo Search Builder.
Con: There are also fan sites for Paris Hilton and serial killers.
If publishers embrace CSEs, then it could decentralize the search experience for consumers, turn CSE publishers into search engines, and create new targeting opportunities for advertisers (e.g., targeting readers of Publisher X who search for Query Y).
This future, if it is ever to exist, will be some years away. Should it ever happen, Google might not have come up with the idea, but will be able to take credit for popularizing it. Innovation isn't everything, after all. Google is simply playing to its strengths; its real innovation is in marketing.