Abbett.org

Dammit, wake up, GM!

I'm delighted to hear that 10,000 Chevrolet Volts will be rolled out this year, but their 20th-century reliance on dealers for sales disappoints me.

Direct quotes from my 2004 Chevrolet purchase:

• "I'm not going to sell you the car you want, I'm going to sell you the car I have on the lot."
• "How much is it? Well, how much do you want to pay?"
• "Let me check with my manager."
• "For only $400 more, you can get clear-coat protectant."

Over 52,000 people (myself included) have signed up on an unofficial Volt waiting list—and GM won't be doing a thing with it.

I doubt every single signatory would be ready to fork over the MSRP on day one, but it represents serious demand. Any behavioral economist would insist that the manufacturer capitalize on their desire for instant gratification, i.e. let your forward-thinking early adopters buy direct.

Think about it, GM – you have serious prospects who have probably been following the progress of the Volt for three years. These are your most enthusiastic grass-roots spokespeople — and you're going to force them through the highly variable customer experience of a Chevrolet dealership? (With all due respect to hardworking Chevy dealers: Apple Stores you're not.)

This has "wet towel" written all over it.

I want to go to a website, pick my Volt's color, click "Buy," and have my new car delivered exhaust-free to my door.

Filter a table client-side with jQuery

My current project at work involves a lot of time-consuming medical data retrieval that makes server-side filtering (by provider, by date, by department, by data type) unpractical. Given that we are forced to load large batches of data up-front, I figured this would be a great opportunity for client-side filtering.

Fortunately, jQuery (version 1.4.2) and a bit of ingenuity made it quite easy.

In the example below, I use e-mail messages as my data elements, and I have three possible filters: year, tag, and sender.

Each filter can be selected using a drop-down:

Upon selection, the e-mail list will automatically update to show only items matching the selected filters.

The trick is this: each row in the table semantically describes its data using the class attribute. For example, a row containing an item from 2009, sent by Sarah, with the tag Inbox will describe itself thusly:

<tr class="year-2009 sent-sarah tag-inbox"> ... </tr>

The filter drop-downs have option values that correspond to those classes, plus a wildcard option with an asterisk as the value. Here's the year filter:

<select class="filter">
  <option value="*">All years</option>
  <option value="year-2010">2010</option>
  <option value="year-2009">2009</option>
  <option value="year-2008">2008</option>
  <option value="year-2007">2007</option>
</select>

Whenever a select element changes, jQuery springs into action. It collects all the non-wildcard filter values that have been chosen, concatenates 'em into a single selector, and then hides the right table rows. Just read along in the code—it's all nicely commented.

• Live Example
• JavaScript code

Usability of Brookline Library website

Update, June 30: Talk about responsive! The library website has already been updated, with hours and book search right there on the homepage. Bravo!

Every few months, I visit the Public Library of Brookline's website. Given my recurrent difficulties, I suspect it may have been designed for or by librarians, rather than patrons.

There are really two things I do on the library website: search for books, and check the hours. Unscientifically, but with my designer's hunch, I'd proffer than 98% of visitors want to do the same.

But how?

First, I see the logo... then I see the welcome image, then I'm drawn into the many links below, hunting for some information scent.

Ah, wait, there's a navigation bar. Maybe the library's books are under "what we have." I click -- nope, just bureaucratic memoranda.

At this point, a bit frustrated, I click back to the home page, and start to scroll.

Ah, a search box! My way out!

I enter the name of my book. Click. My search returned no results.

Only at this point do I notice a tiny collection of secondary navigation links in the upper right corner. Here's one: "Catalog." Younger readers of abbett.org might not be aware that in the olden days, a library's book records were stored in a giant wooden cabinet called a "card catalog." If you're under 18 -- and come on, how many library patrons are minors? -- you're SOL with that link.

"Hours & Contact" is up there too.

This is all a very long winded way of saying: put the most requested resources very conspicuously on your homepage. The top of the homepage should list the library's hours and provide a simple, Amazon-style search box for things in the library's collection (not library website content).

Longer term, the library folks should analyze their web site logs (and search logs!) to find out what are the most frequently sought piece of information -- and make those things the easiest to find.

Good healthcare cheap

I read this and was bummed: the state's cap on Harvard Pilgrim's rate increases have been overturned.

What I find surprising (well, maybe I should stop being surprised by hypocrisy) is that the the Massachusetts Association of Health Plans doesn't want Government dictating their prices, but they're happy to advocate for Government to dictate what health providers can charge.

As I see it, the insurers need to man up if they don't want to be overtaken by a single-payer system. Blue Cross and Harvard Pilgrim should stop whining and just drop Partners.

They don't because there's an insidious behavioral economics problem beneath it all.

We in eastern Massachusetts think it's our right to get the Very Best Healthcare (however that's measured), so our insurers are scared to lose business by not including Man's Greatest Hospitals in their networks.

Of course, we have no evidence that MGH, Brigham, etc., are any better at treating routine issues than suburban non-Partners facilities (if you can find one anymore).

There's a design challenge in here.

1. How can we make it visible to healthcare consumers that they can get a full suite of good healthcare services at much lower prices completely outside the Partners cartel?

2. How can we make such data relevant to individual patients' lives and patterns of care?

3. How do we start before hospitals actually publish comprehensive performance metrics?