In an email conversation, Danny Sullivan (he of Search Engine Watch fame) and I recently were discussing last week’s post on Froogle. Danny disagreed with my premise that Google’s actions were inconsistent, in fact, he believes they may well be consistent with a new and evolving interface approach that he calls “invisible tabs.” He explains the idea here. The gist: search engines will intuit what you are looking for behind the scenes, and deliver to you the results most consistent with that intuition, making the tab format redundant in the first place.
As Danny put it in an email to me:
The real departure is going to be if Google finally makes the jump and gives you back 10 product/Froogle results at some point, and suggest that you might also search the web, for some queries, rather than the web dominance we get now. That will be them fully putting into play this whole invisible tabs concept that I’ve been talking about recently.
Danny points out that Google already does this with News. Try searching for “George Bush,” for example. You’ll see News results at the top. Google is intuiting that you wanted news on George Bush, or at the very least, that news about George Bush is relevant to your search.
Same thing for Froogle results, Danny explains: “They’re hitting the Froogle database in automated fashion, and if the automated system feels confident enough, you get Froogle results displayed. No different really in look, feel and operation than searching for “iraq” and getting news results.”
Well, yes…and no. What I find interesting is this part of the idea: ” If the automated system feels confident enough, you get Froogle results displayed.” No matter what, code = architecture, and architecture = politics. Somebody had to code that Froogle algorithm to determine its confidence/intuition with regard to your search. Google, and any other search engine worth its shareholder’s money, will never tell you how it makes those decisions. They are the Invisible Hands of the automated search process. The men behind the curtains.
And therein lies the interesting bits.
Regardless, we should all give Google a lot of credit for having neither paid inclusion nor referral fees in their shopping engine. That is leaving a lot of money on the table toward a greater end, and an indication of the philosophy which guides them.
My partners and I were just talking about this, in a slightly different context. It is hard to decide when to attempt to intuit intent from a user in code, and when it goes wrong, it goes really, very wrong (think Clippy). Some people don’t like the idea as a matter of policy, which I think is wrong-headed.
I also think people become really enthused about the whole thing and it easily becomes annoying. Google – so far – seems to be striking the right balance, which is extremely impressive.
I think the real value of such measures can only be derived by live testing – out of the 400K people who did that +”George Bush” search, how many hit the “web” button thereafter, how many added a search term and re-searched, etc.
As a side note, I’ve noticed that once in a great while, I get an exit redirect out of Google (meaning the URL returned for a search points to a Google server and they issue a 302 redirect to the desired location) – I’ve verified the statusbar flicker by searching my browser history, and the cached version of the page. I have only seen it twice, but they do appear to be doing some sort of statistical exit sampling. Has anyone else seen this?