I was lucky to sit through this presentation from Doug McClure and his great panel.

| View | Upload your own

In no particular order here are my notes and comments:

1. The Great Unknown – Nobody knows with a great deal of certainty whats going to happen
2. Who Knows, Who Cares – Stop worrying about every little thing you don’t know or can’t know about. Too much time is wasted trying to identify, quantify and mitigate risks with uknowns. If you need to change, then change.
3. Be successful and make money by helping other people make money. Great point. Always be adding value to your customers and you will continue to be successful.
4. Target non-consumption and non-consumers Great reference to Clayton Christensen’s business teachings about providing ‘good enough’ alternatives for consumers who are not open to current offerings because they are priced out of reach.
5. Minimize the chance for competing directly with entrenched players.
6. Question your work regularly Be reflective on the activities you’re doing and the value being added.
7. Read your product out loud. Too often we spend a ton of time trying to make broad sweeping changes when all that is necessary is to make minor textual changes. “Too much focus on pixels, not on words.” “Rewrite First, Redesign second since the design should fit the content”
8. Err on the side of simple. Many projects fail because they try to do too much. The longer it takes to develop the less likely it will launch. Force people to get into the ‘new zone’. People like to work on new things so keep them working on new projects. People lose interest and steam with extended proejcts. Resist the urge to do ‘more’ next time around.
9. Invest in what DOESN’T change now and ten years from now. Invest in key core competencies such as customer service for FedEx and shipping for Amazon. Follow the things that people want
10. Follow the chefs. When you look at the great Chefs, they succeed by giving it away. They give away the recipes and it comes back to them.
11. Interruptions kill productivity. An obvious but great reminder that when you have close colleagues that continuously talk to you, you don’t get much done. Essentially, the little interruptions break the work time down into smaller and smaller bits that make everyone less productive. Passive communication like campfire provides a more formalized way to communicate ‘as available’.
12. Roadmaps send you in the wrong direction. I disagree with the statement but agree with the premise that rolling an 18 month roadmap to customers will ultimately leave people disappointed. Quote: “It’s ok to think about the future just don’t write it down.” This is just too much of a broad statement
13. Do the right thing at the right time. Per the previous, post I agree that you need to make the right decision for the right time
14. Be clear in crisis. Be open and honest with teh public when a mistake is made. “The web doesn’t stop even though you do”
15. Make tiny decisions. “When you make tiny decisions, you can’t make big mistakes.” Celebrate small launches and it is a boost for morale.
16. Make it matter. If something doesn’t matter, don’t do it.

Presentation Key Takeaways:

Modal Dialogs are seductive

1. Provides a way to get around links and the challenge of process flows
2. Developing process flows is hard
3. Pop-ups have been a shortcut to avoid tough process flows
4. Ajax has also become a similar crutch
5. Ajax Design Patterns is a good reference to learn Ajax UI patterns
6. Lightboxes are the equivalent of Pop-Ups 2.0 and are ‘ghetto’
7. You should only use modal dialogs when you want to completely capture a user’s attention

Ajax looks like a performance enhancer

1. Ajax can cause performance problems
2. Always consider performance vs. scalability
3. Not a good solution for performance enhancements

When considering pre-fetching, examine whether it is better to be DHTML or Ajax

1. Examine the percentage of the time the pre-fetched content is actually opened
2. Is there a significant cost associated with retrieving the extra data
3. Examine the current page response time without the hidden panel loaded

Metrics get fooled by Ajax

1. Google analytics is good for the web 1.0 world
2. For slideshare Google analytics inaccurately reported the bounce rate by 400% higher than it actually was
3. While the new Google analytics (currently in Beta) will provide ‘event’ reporting, it is expected to be limited to 100 events per session which is too small

Related content is an opportunity for optimization

1. Many ‘related content’ algorithm’s will pull from tags and it doesn’t necessarily show content that people want
2. Slideshare, targeting more page views, changed the algorithm and increased related content views from 4% to 7%.

Uploading Fishhook Method

1. Slideshare uses a ‘fishhook’ method for uploading
2. When the user selects the file(s) the system immediately begins to upload the files without needing to press the submit button
3. This leverages “invisible Flash” paradigm to do the work behind the scenes
4. Provides user’s with the needed feedback

External Widgets

1. External widgets (add this, gigya) are a great way to get up and running
2. The third parties manage the changing APIs instead of you
3. Users need a direct path to data

SEO and Traffic Nightmares

1. Flash and JavaScript are not SEO friendly
2. Post-fetch data is not seen by google
3. Bake any code or content into a page when loaded if you need google to see it
4. Make sure you get credit traffic as initially S3 was getting credited with some of slideshare’s traffic

Other Points

1. Don’t just save data, provide users with feedback that something has been done
2. Animation: Dial it back
3. If Fonts are important, Flash is still lighter and better since JPEGs are still to heavy
4. Flash Full-Screening is OK as long as it is user opt-in
5. For apps such as the multimedia editing tool, full screen is appropriate
6. Using invisible flash can be great and there is a dojo javascript library to connect to flash local objects

It was a great presentation and you can see the presentation online here

• Consistency for the sake of consistency is not worth it, do it if it benefits the user
• Let users drive changes in product (Adoption of RAW drove Adobe LightRoom)
• Use defined toolkits as much as possible to drive consistency
• Use of 508 compliance (accessibility) to drive unit testing.

Session 3: Web App Autopsy While a few minutes late due to technical difficulties, the session was very enlightening. The panel represented folks from a variety of web-app based businesses including WuFoo, Blinksale, Feedburner, and regonline.

• Each company used a different language to develop their apps
• Since the first version of the apps were released, all estimate they have refactored at least 50% of the code base
• Blinksale commented much of the refactoring was because rails was young
• For Blinksale, the billing portion of the application accounted for 31% of the codebase. It was very complex.
• Feedburner evolved their metrics over time with subscribers and reach (defined as subscribed and opened)
• Support requirements varied widely (8-234 per day) across companies
• Implementation of a knowledgebase reduced support requests by 30%
• One company practice is to reduce support request for a given feature to zero in six months or remove the feature
• All online companies had their worst month in December and three out of four had their best in Jan/Feb
• regonline
  • Split testing has found an 11% increase in registration when a video spokesperson was used
  • Looks for small tweaks to make big differences. A logo that included ‘look behind the screen’ led to a 339% increase in click-throughs
  • Don’t focus on the early adopters but instead
• Wufoo subscription response numbers
  • 7000 Invitations sent
  • 3500 Created an account
  • 50 Became customers with a 50% discount offer
  • Ads on account confirmation page had a 10.5% click-through rate

Session 4: Grids are good While this session was extremely rushed, Khoi Vinh and Mark Bolton did a great job of introducing the audience to the use of grids. I will definitely examine grids in further detail.
Session 5: Ruining the user experienceAaron Gustafson and Sara Nelson did a great job of outlining an approach to stratify interaction design: No Frills, Dress it up (add CSS) and Make it sing (add javascript/ajax). With this approach the designer is able to gradually enhance a page (or element) while ensuring maximum compatibility across browsers and platforms. I like the concept of evaluating to what level of interaction is appropriate.
No Frills

• Make content accessible
• No distractions
• Clean Semantic markup
• Light, fast downloads

Dress it up

• Refined Visual Design
• Simple Interactivity
• Some Flash
• Cross-Browser compatibility
• Styles for alternative media

Make it sing

• Responsive Interface elements
• Predictive data delivery (ajax to reduce errors)
• Allows for more customized interactions

Slides can be found here