So what’s wrong? Take a look:

The list is ordered by when I added the subscription. I can manually change the order, but there is no sorting. Google Reader breaks a cardinal rule of presenting lists. Manual sorting is great, but by default use (or at least support) the most natural dynamic sorting of the information. In the case of Google Reader, I need to see those feeds sorted by unread messages. It’s too bad such a great product has such a lousy problem.

(On the off chance that I just missed Google Reader’s sorting option, please let me know!)

The BBC History site is an excellent place to learn about WWI and WWII.  Among the many great resources are a set of animated maps covering things like the fall of France during WWII, the landings on D-Day,and the Western Front from 1914-1918 (screenshot below).

A static screenshot doesn’t do this tool justice.  As you play it, the map pans and zooms to follow the action, it shows troop movements and major offensives, and it redraws the front line with each major battle.  Why does all this come across so well?  Because the creators of this tool made good trade-offs given that their medium was the web.

Here’s a quick comparison of 3 different mediums:

So if the author were trying to find the right map for a book, they might choose a static but high-resolution, highly detailed one like this WWI map (3.4 MB).  For television, they might choose to animate the map and include voice-overs that tell one specific story-line, but would avoid text as much as possible.

Since in our case, the medium is the web, the author chose a fairly low-resolution map, animated it, including pans and zooms to get more detail when needed (overcoming some of the resolution limitations of the web), included text (because web surfers typically prefer to read rather than listen), and made it interactive (to allow users to follow along at their own pace, taking advantage of one of the key benefits of web content).  Had the author put up a large static map, or an animated, non-interactive movie, it wouldn’t have been nearly as effective.

If anything, the map should be more interactive, allowing the user to pan and zoom himself and click on the timeline as an alternate way to jump around the animation.  And, even though users’ screens are typically low resolution, if this map scaled with the user’s window, it would allow the lucky folks who have high-resolution monitors to take full advantage of their real estate.

If you think about your medium, you have the chance to turn something good into something great.

Things I plan to try from the list are TrialPay and promoting software on one day sale sites, such as BitsDuJour or GiveAwayOfTheDay.

Some of the 100 ideas are specific to downloadable software, and some of them are specific to micro-ISVs. Even if 99 of these 100 ideas are not right for your product, that means one of them is and you have an immediate action you can take to boost sales in 2009. Go check out the list.

• Never show your users error codes - ever
• Pass error information to technical support without user involvement - when acceptable, solve errors proactively. Make it so you can easily correlate the error with the human user that was affected
• Exceptions, stack traces and HTTP error statuses are just another form of error code - keep them tucked away
• Take care in crafting your error messages - not too much or too little information, all the relevant information
• Be friendly, humble and apologetic - you screwed up, not the user
• Make error messages actionable - empower your user with the next step they should take
• If you absolutely must show an error code, tell the user why and what they need to do with it

Provide users with all the relevant information and tell them what action they should be taking next.  Users are usually pretty understanding of errors if you let them know exactly what happened, why it happened, and what they should do about it. The two biggest sources of frustration to a user in the face of an error are, “I don’t know what to do now” or “I would know what to do now but I don’t have the information.”

What I’d like to do now is play a little Extreme Makeover: Error Page Edition.

The first contestant is my friendly neighborhood bank, Wachovia. Nothing like kicking a company while they are down. Like most banks, Wachovia provides online banking services to their customers, and like most banks, these services are essentially web front ends to a mainframe. The thing about mainframes is they predate our online, real-time, 24/7 world, so many mornings when I get up at 4 AM EST, many of the services Wachovia offers to me online are not available. My morning error page comes with an amusing and ever changing variety of error codes. To Wachovia’s credit, these services have a good track record for me of being up by 6 or 7 AM and the English in their error message is rather helpful. By applying the rules above it can be better.

The error page is already friendly and actionable. This is a good first step and is more than we’ve come to expect from most error pages. It would be better if Wachovia wasn’t so stingy with information, if they made it clear why they are providing the error code and what should be done with it, and if they recognized that it is an inconvenience (”this inconvenience”) versus just a potential inconvenience (”any inconvenience”).

The next step for Wachovia would be getting rid of the error code. They presumably provide the error code so the user can pass it along to the customer service representative on the phone (I’ve been asked by Wachovia for the error code before). But it’s a bank and I’m logged in. They know who I am! Their systems should be making that information available to customer service representatives that need it without me needing to relay it to them. Even better, send me an email when the situation is resolved saying, “We’re sorry about the error that occurred while you were on our online banking site this morning, the problem has been fixed and you can access the service now.”

Next I’d like to makeover a company many consider to be the world’s top super model. Apple. Sacrilege? Not everything coming out of Cupertino smells of roses and gumdrops. Take the Time Machine (Mac OS X’s backup system) error dialog that unfortunately every Time Machine user has seen too many times. Apple does a good job following many of our rules, there are no error codes for example. Yet this dialog is immensely frustrating because it provides no meaningful information. What are you supposed to do now?

Making over this dialog is a matter of providing the user with more of the information that Time Machine obviously has, so the user can make an informed decision about what to do next.

The new dialog could be furthered improved with a link to information on how to fix the permission problem.

No matter how polished and foolproof you think your product is (and most of us think our products are mostly bubblegum and bailing wire), there will come a time when a user encounters an error. These are the times that try our product’s soul. The product’s response in the face of errors can make a good product great or a great product crumble.

Butterfly Error Dialog photo by schoschie

If you don’t have time for all 10 (shame on you), the 3 most important are #1 Padded block links, #7 - Auto-focus on input, and #10 - Verbs in labels.

But in enterprise and consumer software, choice usually means “what product level or feature set do I want” and it is typically tied to price:

• Microsoft offers Vista Home Basic ($199.95), Home Premium ($239.95), Business ($299.95), and Ultimate ($319.95)
• Adobe offers their Creative Suite in Design Premium ($1799), Design Standard ($1399), Web Premium ($1699), Web Standard ($999), Production Premium ($1699), and Master Collection ($2499)
• and almost every web host in the world offers something akin to Bronze, Silver, Gold (Platinum, Diamond, Kryptonite, …) service levels, with increasing prices

Does this amount of choice make sense?  There are certainly counter-examples in software.  Apple only offers one client version of OS X, and I think it makes buying OS X easier as a result.  Barry Schwartz talks in depth about the paralysis, disappointment, and even depression that an over-abundance of choice can bring.  In a world where almost everyone’s too busy to decide, it often makes more sense for you to do the choosing for your customers.

So what can you do to provide useful choices to your customers?  Start from the position of your customer’s needs, and not your own.  Intuit does a very good job of providing choices based on their customers’ needs.

Some good reasons to offer choice

• You have several different segments of your customer base who can clearly differentiate themselves.  Intuit offers Quickbooks for business accounting and Quicken for personal accounting.  Very clear.  I know whether I’m running a business or doing my personal finances.
• Different customers need to deploy your software in different ways.  For example, TurboTax is available as an online application, making it very easy to use.  But for customers who don’t feel comfortable putting all their private tax information online, Intuit continues to offer TurboTax as installable software.
• You offer industry-specific solutions, with specific value-added features for those industries.  QuickBooks Premier comes in specific versions for Contractors, Non-Profits, Retailers, etc.

Some bad reasons to offer choice

• You dreamed up features that somebody might want.  You did this without talking to customers and users.  Rather than understanding your customers’ needs up front, you ended up using multiple versions and add-ons to make your customers choose when they buy–  “We can’t figure out why you’d want all these features, but maybe you can”.
• The internal organization of your sales, marketing, and/or development groups has caused you to create optional add-ons when a customer would best be served by a single product. 
  
• Your competitor offers 4 levels so you think you should too.
• You want to claim to have the lowest prices in your market, so you build a version of your product that’s so stripped down it’s useless.

Why does pricing matter so much, especially these days?

• In a down economy, people tighten their belts.  Spending they might not have given a second thought to gets re-evaluated.
• People will ensure they are getting real value and will decide, usually emotionally, if they are paying the right price for that value.  This doesn’t mean that all prices need to be slashed, but pricing, and especially the value of your product must be made crisp and clear to your customers.
• People that can’t afford the “premium” option will often choose not to buy at all rather than to buy the inferior option. In providing a “status purchase” on the high end you are eroding the status of the normal purchase.

Toilet paper photo by anyjazz65

“Two groups of men were asked to rate the experience of having a diagnostic colonoscopy.”

Whenever you can get product design insight that starts with a sentence like that, you have my attention. And what is the advice? It’s called the Peak-End Rule. No matter how great or bad someone’s experience is, how the experience ends is going to make a very big impact on how they feel about it. Read that last sentence one more time and think about your own products for a minute.

How often do you think about how a user finishes using your product? What’s the last thing they do before they stop using it? How can you make that last thing better? How can you leave your users with a big smile?

These are good questions and I had never thought to ask them. Dan Willis works for Sapient and writes a blog, UX Crank, full of good advice about matters of user experience. You should read it, I highly recommended it.

The Law of Overexposed Abstractions - the users of your software don’t benefit from the same abstractions that benefited you in building the software

Read the full article. It’s gotten some great comments.

Here is an innovation I can get behind. Let’s start teaching kids at an early age that ALL maps are fictions. That the only accurate representation of something is the thing itself (leaving aside Kantian epistemological concerns until at least High School). If the only maps we grow up seeing are the same old projections of geographically focused maps then everything else looks both foreign and wrong.

Here are two maps that show results of the recent U.S. presidential election. Both show how Obama won the urban centers and McCain won the surrounding rural areas. The map on the right has made size a function of population, not geography, and it uses shades between red and blue. The one on the right, from Physicist Mark Newman, is the more informative and innovative of the two. Are we open enough to cartographic innovations? Why didn’t I see something this informative on CNN, ABC, NBC or CBS on election night?

Get the complete story on Mark Newman’s new electoral map in this short video.

And if you then enter your home address you get a map of your polling location. This is brilliant. When you can correctly guess what your users are looking for, you have the chance to delight them. A few things to keep in mind when guessing what your users want:

• Give them what they want, not what you want them to want
• Guessing works best when you know most of your users are in a certain frame of mind (e.g. voting on election day, preparing taxes in April)
• Guessing can also work well if you deliver personalized experiences to your users (Amazon does a great job of this)
• Since guesses can be wrong, don’t lock your users into your guess. Google still returns standard results in the example above, letting users easily ignore their guess.

So while you have to be careful, if you find an opportunity to make a good guess about what your users want, surprise them with excellent service.