08 Jan 2003 | Darrel said...

Looks nice. Let us know how sales go. ;o)

BTW, your 'click to close window' cursor on the sample page pop-ups goes into fits of spasms in IE5.5/PC (the flashes the tool tip on and off continually).

08 Jan 2003 | paul said...

MS must have fixed that for ie6, bc it doesn't happen for me. my cursor changes to an up/down arrow from the normal pointer though onmouseover of the img in the popup, which is kind of odd though.

08 Jan 2003 | JF said...

BTW: When you purchase a report, your name (or company) comes on the cover!

08 Jan 2003 | dmr said...

Does the report include some practical ways to make my own search more 'human'? I realize it's probably non-technical and conceptual, but what is the next step after making the realizations and wanting a better engine?

08 Jan 2003 | SU said...

Does the report include some practical ways to make my own search more 'human'?

The report includes a Best Practices section which brings together 20+ top solutions these sites are applying to each of the six areas tested. And although they're not technical, back-end solutions you can graft onto your site today (sites differ so much in that respect anyhow), they should serve as inspiration for your own search logic and customer experience design.

The bulk of the report shows 25 popular sites doing things right... And sometimes very wrong. We encourage readers of this report to take a look at both ends of the spectrum, get some ideas, visit the sites, and even run a few tests of their own.

We were pretty amazed at how poorly some sites like Apple and eBay fared when it came to basic, human search queries.

08 Jan 2003 | dmr said...

I'm sure results depend on the engine, most of these were homegrown by the companies in question, corrrect? Were any conclusions reached about boxed search solutions that fared better than most?

I'm not trying to extract information for free, just trying to figure out if I take the recommendations and go modify my PHP/MySQL search code, or does the 37 crew have recommendations for out-of-the-box search solutions that real people can use with good results. It would be nice to achieve some good search capability without hiring 17 programmers to rewrite Google every time we want searches in human terms.

08 Jan 2003 | Darrel said...

Note that sometimes there's the opposite problem, where experienced searchers who are used to boolean logic and find poor results when using a 'human' search string parser.

We worked with a product called Autonomy on a large project. They boasted how great their real-language search algorhythms were. The idea would be for you to type in a few sentences in english and it would come back with excellent results. It worked, but we couldn't figure out how to convince people to type in a few sentences when they were used to some simple boolean keyword phrases.

I realize that's not what your report is about...as it is really about just making whatever search tool you are using respond to more different human variables...just thought I'd throw that out as part of the search-engine discussion.

08 Jan 2003 | Steven Garrity said...

We had a very hard time finding quality search tools for our sites. We've finally resorted to using a custom solution one of our programmers put together. Our older sites had used Verity's search with Cold Fusion, which didn't work well at all.

I'm only half-way through the report - but it looks great so far.

08 Jan 2003 | Bill said...

So, why did Jason Kottke get a pre-release copy? Advertising?

08 Jan 2003 | Darrel said...

I'd say getting Kottke to say something is a pretty cheap way to spread the word.

08 Jan 2003 | Steven Garrity said...

Where does it say Kottke got a pre-release? The report was available on the 37signals site on Monday.

08 Jan 2003 | ~bc said...

I'd be more than happy to provide feedback. If you want it, email me a link. No need to specially type my name across the front though.

09 Jan 2003 | Joe said...

Where does it say Kottke got a pre-release? The report was available on the 37signals site on Monday.

Look at the bottom of the sample pages.

09 Jan 2003 | Uh-oh said...

oops!

09 Jan 2003 | SU said...

Look at the bottom of the sample pages.

Jason (and others) provided valuable feedback for us as we worked to refine the structure and wording of this report. We wanted to make it the best report we could with the data we'd gathered. Outside readers helped shed light on things we as authors were too close to see.

Just as you'd hire outside consultants from time to time in order to get a new perspective on a project you're creating, we requested outside opinions on our report. His decision to link to it was his and his alone.

09 Jan 2003 | JF said...

BTW: We uploaded final release screenshots now. Our mistake for publishing the pre-release versions.

09 Jan 2003 | SU said...

To ~bc and Darrel... I misread your earlier comments above and replied sarcastically. I've removed my offending comment apologies to you both.

09 Jan 2003 | Bill said...

Having been to 37 signals contingency design workshop and see a little of the report there, I was anxiously awaiting the report, especially because we are undergoing a huge search redesign project at work. We purchased the report yesterday, and it has already helped a great deal in regards to our project.

Outside of traditional usability testing and user research we have done, it is becoming our best tool within the redesign project.

Just wanted to say it was well worth the hundred dollar cost.

09 Jan 2003 | Smiler said...

Bill,

Surely you used the SVN coupon and saved yourself $10 - hence only $90 cost!