One of the hardest things for a start-up to decide is "When is my product done?" I've worked in many start-ups, and I am advising several start-ups now. Last week after my article about the Minimum Viable Concept, this issue came to a head at one company. The challenge of when to open the website or let the world through the door is a monster that many teams wrestle with. Paralysis by analysis can lead to missing a market opportunity. But, can you get another chance to make a first impression if your product is truly bad?

Adam Seifer, co-founder and former CEO of, one of the oldest and most popular photo sharing sites on the net, said: "I frequently find myself trying to convince partners, advisees, etc., that one of the biggest risks a start-up has is to not launch anything at all—to get so caught up in talking about what you're going to launch and so fixated on details that it feels like you're making progress when instead what you're really doing is moving asymptotically closer to something that doesn't ultimately matter as much as you think it does."

Some companies have truly embraced the world of continuous testing, also called "using your customers as test subjects." Google's Gmail was listed as a "beta" or test version for about five years. Apple's "Siri" voice service is still considered a "beta" (and given that it works for me about half the time I'd say that is an accurate label.) But are these kind of tests for everyone?

"Not so fast," said Art Chang, CEO and Founder of Tipping Point Partners, a New York City-based "institutional entrepreneur" that serves as a seed investor and founder of software start-ups. "Siri is a 'feature' on the iPhone, but the iPhone works. Gmail is an extension of the Google brand, and though it started a little shaky, by the time it came out of beta, it was solid. If you're an established brand you can experiment, but for a traditional start-up, that first product represents a brand promise. So, it may be better to reduce features and functions to have a product that works really well, and does what it is supposed to do. Twitter and Evernote both started as very simple products, and expanded after they got the core right."

Seifer told me about a current project he's working on, launching "maybe next week. The team was really stuck for a long time noodling around on the same minor things and having the same hypothetical discussions. After I pushed and pushed to focus and get something complete—even though it's a simplified version of what we hope to become in the near future—it had a dramatic effect on our productivity. Having a first version of a site live and real allowed us all to buckle down, stop dealing in hypotheticals and stop deferring decisions, and possibly most importantly, start getting real feedback from real users and outsiders."

John Borthwick, CEO and Co-Founder of  Betaworks, a kind of incubator/studio that builds and invest in companies, said "The whole idea of 'done' is a concept that we're getting rid of, because nothing is ever done. The name Betaworks is a play on 'a factory building stuff' but also emphasizes that Betas work. The world is about continually improving and evolving. I remember what Brian Eno said  (in Wired in 1995) about unfinished products , and how you're constantly in the creation process. Before the user participates the product isn't done—they make it what it is. This old view that you create, publish and you're done, is outdated. You put something out, then users arrive, they comment and your thing morphs into something else, then you update and others engage—done is never."

Borthwick believes we are now learning how to live with unfinished products as a society. "The process of marketing and designing and getting feedback is changing. We're entering into a different era of what it means to be done. Examples include everything from Trader Joe's to Nike Customized shoes to fashion companies condensing cycle to idea to design to in store in a very short time. "

These entrepreneurs and investors believe there are ways to figure out a core of a product, launch it, and constantly iterate after that. Is this easier for software than hard goods? Initially I thought so, but then I remembered when razors had 1 blade instead of 5. Is it true that nothing is done until someone else tries it? When are you done? What helps you make that decision? Let us know in the comments.