Most ideas are bad, so embrace ‘continuous discovery’: Marty Cagan

Once an early-stage company finds its business model, the push for customer acquisition often turns its team toward optimization, losing its “innovation” focus. So warns Marty Cagan, the latest speaker who Etsy’s Code as Craft crew has brought to Dumbo to share thoughts on better dev work.

“Once you get to product-market fit, you can’t stop discovering,” he said.

Cagan has worked his way through major tech names like HP, Netscape and Ebay, and now takes part in the tech world as an investor and adviser. As a self-described fan of ecommerce giant eBay, he said that he’s been familiar with the team since 2008 and has known that the idea of “continuous deployment,” a dev method Etsy employs, has long been a team focus there.

“Our job as a product team in a technology organization is to consistently deliver value for the company,” Cagan said at the beginning of his talk. Ideation should move quickly to prototype, then iteration, he said.

The full talk should be viewable soon on Etsy’s Livestream. Check out Cagan’s book, Inspired: How To Create Products Customers Love.

However, more often than not it, he said, it goes like this:

Idea from an exec –> Business Case by an exec –> Roadmap for building it –> Requirements estimation –> Design –> Build –> Test –> Deployment.

Each of these steps yields artifacts, most of which are just wasted time. It also locates the ideas a long way from the people working the most closely with the technology.

The two inconvenient truths of products, Cagan said, are these:

1. Most ideas are bad (either customers don’t care or the design is lousy), and
2. It’s going to take three to four iterations of a product before you see time-to-money.

That’s why the sheer number of steps in the above is so problematic. Instead, Cagan argued for an approach he called “Continuous Discovery.”

Effectively, he argued for loads of prototyping. He told the story of Google Glass, a hardware effort, that built stacks and stacks of prototypes. He emphasized that these prototypes were seldom functional. For software teams, he said that prototypes should hardly ever rise to the level of production code. More, what he called, a “live data prototype.”

He said teams should be rolling out 15 iterations per week.

Prototypes are critical, he said, because “none of us know what we want from software products until we actually see it.” This squares with what the product team from Huge said when we visited there.

Cagan was a big booster of designers too. Companies, he said, don’t understand design. They think it’s look and feel. When, in reality, it’s the whole experience of the product and coming up with the right experience can be even harder than engineering.

So he strongly advocates co-location of designers, product managers and devs in the same physical space.

“Innovation is directly related to the number of attempts,” Cagan said of the number of actual deployments. That’s why shortening the time from ideation to deployment is so key, he said.

On the growing case for the New York tech community, Cagan said of the dev talent here, “I honestly cannot tell the difference between a team in New York and a team in Silicon Valley.”