Iterative Product Development to Enable Breakthroughs: NPD at Google
Today at the Product Development Management Association's Annual Conference, Marissa Mayer described Google's unique use of iterative product development to enable breakthrough products.
She described the difference between Google's approach to innovation, and the traditional "castle-building" approach. Castle-builders close the curtains, go dark for awhile and then release a produce when it is perfect (or nearly so). Google, on the other hand, will often release features very early, then rapidly iterate the product based on user input until it is perfect enough.
As a Web service, Google is uniquely positioned to make the most of this strategy. Ms. Mayer described Google's sophisticated system for tracking usage data to provide hard evidence to answer such questions as "Which user interface is more effective?" that help them identify and fix problems. They can set their servers to redirect only a small number of users to a test site for a new version, and when they want to roll to the next one, it is as simple as rolling code into their production environment, a process they have honed.
One criticism of such an approach is that it leads to only incremental innovation. Ms. Mayer showed how Google uses this to lower the hurdle for riskier, breakthrough products. If something doesn't work, they can just change it or kill it with little risk to the rest of their business. With a little creative thinking, other services companies can directly leverage the central principle to release imperfect products to small numbers of customers early and then perfect them with direct user experience input over time.
This is obviously much harder to do when a product consists of much more than lines of code on a server. However, creative use of rapid prototyping technologies may make it easier for us to get user feedback earlier in the process than we do now, which will make the development and launch processes go more smoothly.
How early do your customers start using your new products? How can you get them to do it earlier - maybe much earlier than you do now?
She described the difference between Google's approach to innovation, and the traditional "castle-building" approach. Castle-builders close the curtains, go dark for awhile and then release a produce when it is perfect (or nearly so). Google, on the other hand, will often release features very early, then rapidly iterate the product based on user input until it is perfect enough.
As a Web service, Google is uniquely positioned to make the most of this strategy. Ms. Mayer described Google's sophisticated system for tracking usage data to provide hard evidence to answer such questions as "Which user interface is more effective?" that help them identify and fix problems. They can set their servers to redirect only a small number of users to a test site for a new version, and when they want to roll to the next one, it is as simple as rolling code into their production environment, a process they have honed.
One criticism of such an approach is that it leads to only incremental innovation. Ms. Mayer showed how Google uses this to lower the hurdle for riskier, breakthrough products. If something doesn't work, they can just change it or kill it with little risk to the rest of their business. With a little creative thinking, other services companies can directly leverage the central principle to release imperfect products to small numbers of customers early and then perfect them with direct user experience input over time.
This is obviously much harder to do when a product consists of much more than lines of code on a server. However, creative use of rapid prototyping technologies may make it easier for us to get user feedback earlier in the process than we do now, which will make the development and launch processes go more smoothly.
How early do your customers start using your new products? How can you get them to do it earlier - maybe much earlier than you do now?
posted by Katherine Radeka
at
8:54 AM
0 Comments:
Post a Comment
<< Home