As business moves on from the manufacturing of physical goods, our products are now represented as collections of information and a user interface we design. This is our “one” output, and it must suit all our users and evolve over time. Or does it?
How can we balance the “need” for change against the needs of our customers?
The automotive industry has been a major technology driver throughout the 20th and into the 21st century. Progressing from powered carriages (and even racing sleighs) to modern technology showcases on wheels such as the Toyota Prius and Chevy Volt, each model year brought something new, something different.
Unfortunately, technological innovation doesn’t run on a neat yearly or quarterly schedule of change. Pressed by marketing departments for a new model to showcase, often the only difference between model years was a tweak of the trim or subtle change in design.
Detroit automakers are particularly prone to this behavior, as the only real difference between a Chevy Silverado and GMC Sierra is the name stuck on the outside. The staff employed to differentiate the two cousin vehicles has plenty of spare time to create new “improvements” for each model year.
The other option of course is to only change what is necessary and eschew yearly updates. Two of the most famous vehicles ever produced followed this pattern with great success. Both the Model T Ford and the Volkswagen Beetle remained essentially unchanged for their production lives. The Model T was produced for 19 years while the Beetle was survived for an astounding 65 years of production in various countries.
Each car was designed to be “transportation for the masses” and functioned quite adequately, even as competitors eclipsed the models in all measurable ways.
Yet these cars still held something special, something upon which strong followings are built. Ford continued producing the Model T’s engine up until 1941 because people simply wouldn’t let go; maintaining and upgrading Volkswagen Beetles is still a major industry world wide.
But this story isn’t about cars, it’s about the internet. If you do business on the internet, your webpage is your one face. At eBay, our site and the guts behind it is our one “product”, the one product all our users see and use. So how do we balance our urge to upgrade against disrupting the natural flow of business with a new interface?
In the automotive world, you can produce a new car while continuing to support the older model, because you are shipping thousands, if not hundreds of thousands of new vehicles. Those who want to upgrade will grab the new shiny and reap the marginal increase in rewards; those who are content with their transportation will stay put and change the oil as they always have.
Not so in the internet world, when you upgrade the site, it’s upgraded for all your users, whether they like it or not. We push everyone into a brand-new Pontiac regardless if they were happy in their Beetle.
Craigslist has followed the “change the oil” pattern very well. While it isn’t pretty, or as “full featured” as modern competitors, Craigslist holds its own just as the Beetle defied the market. It’s simple and effective, enough said.
And if we still feel the need to change? We should take a page from the auto industry: standardize and differentiate.
At eBay, we’ve build a robust set of standard APIs so developer can write an application on version “440″ of our API, and it will continue to work as long as eBay acknowledges that 440 is an acceptable interaction model. At any given time we may support 20 or 30 versions of our API. We can “afford” to keep these models active because we have modularized our data away from the interface. The pieces “just work”.
This is where our 3rd-party developers come in, creating different versions for different people. Some individuals want to see eBay Like It’s 1999, and some are looking for the screaming edge of cool. As long as these applications respect the modular APIs, each user can have their own version that suits them best.
Now, we’re still a long way from total success, our primary interface is still intimately tied to our databases and this is the version that all our customers must use, for better or for worse.
As the web evolves, I see the ability to add, remix, and recreate our experience of it to be a fundamental right, as essential as being able to read the last chapter first, if we want to. Opening up the doors to your data embraces our profound desire to make an experience ours, directing our future as we see fit.
