I've been listening to the audiobook of 10 Faces of Innovation by Tom Kelley of IDEO. He has a great insight into something I sorta knew, but had never thought about why it was so.
The basic idea is that showing just one prototype to a customer is bad. This tends to force them into a binary/reactive decision, one that is often determined more by their relationship to the creator of the prototype than to the item itself. Given multiple items to evaluate, the attention turns more to the items, and more feedback can be gathered, since it won't be perceived as direct criticism.
The telling analogy Kelley provides is that of a husband that is told by his wife that she has purchased a new dress. She then tries it and asks how it looks. Of course, it is difficult for the husband to say anything negative, because the question is really about how she looks wearing that dress. However, given multiple options to consider, the husband is actually more free to give honest feedback.
Now, it can be expensive to buy multiple dresses, or build multiple prototypes. The answer then is to use lo-fi prototypes: whether it be pictures in a magazine or crude drawings. Still, the more prototypes that are created, the less pressure there is to get it right the first time, which is a crushing force against innovation.
Having heard his brother speak at a conference in 2006, I love what IDEO are doing. If didn't need to make quite so much money, and had any confidence in my abilities, I would like to work there...just to live at that creative level every day would make me so happy.
Wednesday, May 30, 2007
Thursday, May 24, 2007
I've had this in draft for a while, can't figure out what the point was, but I don't want to think about it anymore. I've seen so many people chafing about the enterprise IT shop at one of my clients...unfortunately only I can do is add another forlorn whine to the choir.
My friend sixty4bit recently blogged about some trust and power issues at a client we have both worked for. Their behavior is bureaucratic in the extreme, they run all projects through a slow and cantankerous review board for all of their technology selections. In the end, the review board exists because the organization does not trust software developers and does not want to empower them to make decisions. The end result is that they slow everyone down and drastically reduce the effectiveness of the organization. The overall atmosphere created is one of fear of anything new, where people stop asking to try new things, because it isn't worth the time to get them approved. In a lot of cases, you have to get approval just to try something.
Laughably, this client is attempting to move to agile development processes, which are certainly focused on empowering the people who are doing the work to make decisions (although that is more of a Lean idea), while continuing on the opposite track in terms of empowering individual projects to meet their customers' needs.
In reality, particular technical decisions about technology selection can be contentious, especially in regards to projects across the enterprise choosing incompatible technology stacks. There is also a need to coordinate purchases across projects to gain negotiation leverage. Even more important, although completely overlooked by these bozos is that the license terms for all software need to be read by actual lawyers. However, it's clear that the real issue here is one of trust and power, tied up with the fun of getting to decide how to spend other people's money.
I am working on a devious plan to: a) get them to trust me, b) keep them feeling powerful, and c) allow for innovation and progress. Reading "Influence: Science and Practice" by Robert Cialdini tonight. Maybe that will help with a and b.
Posted by Matt McKnight at 12:03 AM
Thursday, May 17, 2007
Pretty big round for uLocate. Famous for providing support for Flagr, the mobile version of MapQuest etc.
I tried to download the new Where (how much did that domain name cost?) app to my phone, but it's not supported as of yet...
Their mobile technology is sorta cool- going for the whole write once run anywhere thing for mobile apps with an xml style application definition. A location specific version of something like mFoundry. Of course, when they don't support your phone, it's a little harder to buy into that concept.
Anyway, XML as a programming language? Just because it worked sorta okay for HTML doesn't mean there aren't better ways to do it- especially when you are talking applications.