I am not envious of whoever has the job of doing Ubiquity’s new logo. (if you didn’t know our current name and logo is a little too close to Ubisoft’s ) So why do I not care to be that designer? Take a look at some common logos in the Web 2.0 style. It’s not that I particularly like the older logos as they were, it’s that the Web 2.0 makeover didn’t add to them, they were either just as bad or worse.
Web 2.0 design largely consists of soft edges and colors with a combination of clever naming/typos/url’s, dots, and shapes.
The only logo that I think improved was Craig’s List:
vs.
I think Web 2.0 style is largely rejection of early 90’s web design, which wasn’t much of design and more of barely keeping color consistency. It is much the same as what happened in the 60’s when design movements were enabled by cheaper color printing and the ability to do spot color matching thanks to the Pantone system. Notice how everything from the 50’s is dark green, corn yellow, or puke grown? It’s because matching colors across 2,000 stadium seats, or even just living room chairs was hard. It was solving those technological problem that made the 60’s and 70’s so colorful -otherwise acid wouldn’t have been as much fun ; )
Anyway, this is supposed to be a short post, lets see what Web 2.0 doesn’t fix:
vs.
The problem with IBM’s logo is the 80’s. Web 2.0 didn’t do anything to the logo fundamentally, it just added a reflection (a common trick Apple made popular for generic media to make it stand out) and add a bezel. IBM’s logo is too fucking busy, it is too overbearing, its ugly to its core, and no amount of window dressing changes that.
Then we have AT&T:
vs.
Uh oh, something went right here, maybe it was because the logo revision was a paid project done by professional designers. But I don’t think this is Web 2.0, how many web companies have a 3d logo? What improved the logo was the font change over is taken from the Swedish 60’s + 70’s type revolution, not Web 2.0. Yes, it is in lowercase, but the fundamental change is the well-balanced type used by the Swedish Typographers. If you think I am splitting hair here let’s look at Apple’s conversion:
vs.
You just ruined the power of the Apple logo. Killed it, hipsterized it. Regulated it to a small company, trying to be cool but not. “But wait,” you may say, “Isn’t Apple’s logo Web 2.0 itself? Didn’t Apple originate the reflection and drop shadow, and so many other Web 2.0 styles? You can’t say that Web 2.0 is bad if the original logo is already Web 2.0- just this one design!”
To a degree you are right, but you know what? That isn’t the Apple logo, and neither are these:

In my last blog post I lamented the time it takes to produce clip-reels.  Instead of bashing clip reels usefulness, however, I am determined to make their value proposals outweigh the waste involved in creating them.  To do that I created yet another value state map, this time of the "production side" of creating video.

So what doesn’t add value to this equation?
• Watching the video definitely adds value to the researcher.  Anyone that doesn’t enjoy that part shouldn’t be doing usability research.
• Reviewing video and reviewing notes, I hate this part.  It takes forever, requires massive amounts of attention, and I tended to do it several times because I didn’t know what I was doing the first few rounds.  But it has to be done by someone, if you have an assistant they can do it while the study is going on, but it still requires man hours.
• Formatting the commentary and syncing it between blog, wiki, and video comments is the real reason I hate step two.  If a team member adds their opinions, attaches something to a bug, or edits comments I then manually re-sync each site.  its hell, and the action of re-syncing each site doesn’t add value to the changes, its just labor that should be automated.
• Converting and uploading video is a pain.  Most of the optimizations are done or waiting for technology to advance.
• Tagging the bugs adds value, the time taken to correlate the video, Get Satisfaction, and the bug itself do not.  It requires a lot of concerted effort
• Blog posts, maybe its being a Mac user, or a titch of OCD, but it takes me well over 4 hours per blog post.
So a quick count reveals most of the waste lies in syncing information between different steps.  There are also some optimizations to be done by moving video conversion and uploading to more automated batch processing.

1. Review and tag video once.  iMovie 08 (while unable to anonymize video ) has tagging but lacks the ability to export those tags or chaptering.  So while it makes clip reels very easy it will take some hack to make that information flow.
2. Have that information flow between steps by automatically converting to separate sources and combining video source with tag information when possible.
3. Use powerful video converter to batch convert video into multiple formats, use of GPU and distributed computing is a plus.
4. Trigger script to automatically upload that video.
That is about as far as we can go with available tools.  The heavy lifting of actual testing and reviewing data is going to be there.  Distributing reviewing that data among multiple people is much harder, for now.  Daydreaming is the third step in value stream mapping, where you are, where you are about to go, and where you could go.  Take the next step, and repeat.  Where we were was a an average combined 5.5-13+ man hours of time per video.  Video that was pushed onto developers, our customers, at inconvenient times after massive amounts of effort.
In the future, a collaboration engine could enable us to tag the video so comments and data are synced not just between steps.  This step would make the process fully open and free.  And the ability for developers to explore through the tagging engine and create their own clip reels?  F’ing priceless:

In my attempt at increasing the usefulness of the usability testing I have been trying to reduce the time and effort required for the programmers and UI Engineers. After some experimenting with different technologies I did some value stream mapping, another technique used by Toyota in its Lean manufacturing system to maximize the value of its products.

Toyota attempts to see processes from the consumer and steps backwards to its raw materials.  It attempts to identify each step of every process as either adding value or waste.  Waste is defined in many ways, but we won’t get into the fine details here- waste is whatever does not add value.  As Shigeo Shingo observed, "only the last turn of a bolt that tightens it - the rest is just movement."

Here is a current state map stepping from the developer finding out about the video’s to their optional contributory actions:

Chance of getting anyone to actually pay attention and not multitask, none. 

Lets break it down and define what parts are adding value and what parts are waste:
• Personal emails do not add value, it takes time and is pushing information on the developer instead of creating a need, or pulling information.  We won’t get into the push/pull philosophy of Lean but lets just say it’s important to create demand, not dump it.  So let’s get rid of that by combine RSS and video podcasting to mailing list announcements.
• *The blog post is nice because it provides some context, but blog posts work better if they are topic, not video specific.  If the video is fresh (we’ll get to that later) it takes longer to get the video up there.  So lets move that information to comments and chapter markings.
• *Vimeo is nice and all, but the comments are not inline with embedded video, chapter support is weak, and it requires people to go to Vimeo while on their computer -encouraging multitasking.  Adding podcasts will increase the pull aspect of the system.
• *We can also cut down on time spent watching the video.  Voice is still comprehensible at 2x and following actions at 3x is still possible. Unfortunately, none of the online video services offer this feature inline. That doesn’t mean we can’t provide a 2x video in addition to the normal video.
• If the developer chooses to blog, contribute comments, or write to the wiki that process should be made as slick as possible.  Part of moving the wiki to the Mozilla wiki and off of my site is to use a single sign on for all developers, hopefully they will already be signed on or have their password saved.s  A WYSIWYG editor would be nice as well, but that goes beyond what we have power over, and will be saved for future ideas.

Now we have the makings of a future state map:


50% improvement on minimum commitment is a lot of progress.  On average, however, developers will be spending 45 watching video, reading the wiki, and browsing comments for each of the 3-6 videos per test.  Nor does that account for switching between the normal speed and 2x speed. In reality we improved things by about 25% - 50% for the hard core usability nerd but not enough for the overly worked programmer.  What else could we do in the near-term given the tools we have to work with?

An obvious one is attaching clip reels to bug reports.  The developers are going over bug reports anyway so we get findability and buy-in for free, assuming we are able to put videos inline with the bug reports themselves.  Being short form also combats the tendency to multitask.  Cross referencing with the bug reports adds value to all stake holders by helping determine how critical the bug is.



Clip reels are often thought of as the perfect way to convey how a user is having trouble. In the Usability Testing Handbook Dana Chrisnell warns that clip reels can take 60+ minutes per minute of video. I balked at this notion, surely this was based on experiences with older non-linear systems.  Even using iMovie, however, finding clips across sessions (now which testers did that and when?) putting in commentary, converting, uploading, and making the web page easily adds up to a half or full hour per minute.  Making the clip reels has forced me to do summaries, which are often much faster for both me and the reader.
The drawbacks to making clip reels do not negate the value they can add. They are, of course, no substitute for the full experience. Context is a major part of the story when a user falters. However, the time issues only highlight waste on the other end of the system. In my next post I will examine the production of the video by value mapping it.