how long did FriendFeed take? - Jamie

I'm guessing that the iPhone was under development for at least several years before 1.0 was released. FriendFeed is less than a year old -- I bet that it will be a lot better when it is two years old. - Paul Buchheit

But as you know well, Gmail have got quite many updates after that (but most of them could be counted as minor fixes probably). Just saying that product is good only if it's been maintained. ;) - Daniel Schildt

Does good mean, good enough to go live? I'm a little confused by how good is evaluated. Our hotels meta-search product (wego.com/hotels) has been around for about 2 years now, but we did a re-write a couple of months and it's much much better than before, but still don't know whether it compares to GMail's 2004 level of goodness, or Google's 1999 level of goodness. - arunthampi

Yes, of course good products must continue to improve. The Google of 1999 wasn't very good by today's standards either. It only searched about 70 million pages, for example (vs maybe 20 billion now). - Paul Buchheit

arunthampi, I'm thinking of products that will stand out and have significant impact on the world. - Paul Buchheit

FriendFeed have been under development for just little bit of time but considering that amount it has gone greatly forward. I wish it will develop to even more advanced (but still usable) tool for keeping persons updated without too much of information flow. I wish there would be more features for filtering of information that user could set and control from settings. - Daniel Schildt

2-3 years. Based on Moore's law http://en.wikipedia.org/wiki... - Igor Poltavskiy

I started eventseer.net 9 years ago, but it wasn´t until recently it became good (after Thomas took over 3 years ago). - Amund Tveit

Habari has been in development for about 18 months. It's a "good product" now (though with pretty low market penetration :) but it will be a lot better in a year. - Michael C. Harris

I think the iPhone was under (serious) development for 3 years prior to its launch last June. - Jamie

@Paul, ah in terms of significant impact, maybe we don't measure up (as yet) but im guessing that will happen very soon :) - arunthampi

It depends on what you mean by "good product". Good for lead users? Good for a distinct social milieu? Good for the masses? Also I believe it is very hard to tell when a product has been finished. Gmail 2010 won't be the same as Gmail 2008. - Benedikt Koehler

"Good to use" is not universal since different people have different need and for some of them, tool can be always difficult to use. - Daniel Schildt

Software is generally asymptotic to good, (isn't it?) because as software gets better it attracts more users, each of whom has a slightly different definition of "good". But some more random examples: Windows 1 -> Windows 3 == 5 years; Unix 1969 -> 1985 (by which time it was clear it was dominating workstations except for DEC); Linux on the desktop 1991 -> 2007 (Ubuntu 7); Mozilla/Firefox 1998->2005. It's pretty clear that web software matures quicker than desktop. - Nick Lothian

if you write software you should consider your users as BORC, not those weak dumb individualistic humans. that way you write software that fulfills needs. the more streamlined your BORG can do it the better. leave the task to make pretty buttons to the designers. the last step would be to imagine a human again and make your software human/error-proof. - Chris Hofmann

It looks like the incandescent light bulb was around for 75 years before it was commercialize-able: http://en.wikipedia.org/wiki... - Clare Dibble

Game development is about 2-3 years but not sure if there's much correlation of development time to quality. - Chris Bentzel

Human brain software takes 2–3 years before it is able to reasonably control its host human. - Amit Patel

how long does it take for a good product to go bad or rot? Why do some products still remain relevant after many years? - Shakeel Mahate

The first fax machine was sold in 1861. - Seth

I heard Writely was built and sold (to Google) in 10 months - Stewart Rogers

Interesting question, but I can't answer it because I can't figure out how long anything has been in development other than by the community of people who are working on it, but then I can't correct for the bakedness of ideas when a community starts. And then I'm not sure which of the ideas in the bundle that is a product was really important to its success. Hmm. - Daniel Dulitz

I think it's a question that can be answered historically, if you look at the time from when a group of people commit themselves to the realization of some idea, to the time when a similar group with a similar idea make an impact on the world. The answer for software *seems* to be 2-3 years, much of which is spent fumbling around in an unfamiliar space. For new drugs or new modes of transportation or new paradigms in visual art or whatever the time is different, and there's a lot of variance in any case. - ⓞnor

based on the startups i've seen over the years, 1 yr to prove the concept and 2 yrs to fully launch first "stable" iteration, then improvements from there on out to succeed (definition of success differs based on model/plan) - that'd be my swag of it... - mike "glemak" dunn

I like this topic Paul -- don't see it delved into all that much, at least in places that I tend to keep my eyes on - Eric Berlin

great topic.. longevity is key to successful product. if you are working on something new (not a copy of something else), you got to factor in some iteration time as well. It's interesting that many here are coming out at the same 2-3 year time frame - i agree also - Travis Parsons

I've been PM for an enterprise software suite for 8+ years now and it's taken that long for it to mature to the point where large corporate customers around the globe are deploying it for mission critical applications. We've gone through 4 major releases and countless minor/dot releases in that time. This is technology that originally was developed by a startup back in the 1996-2000 timeframe that was bought and sold via M&A activity 3 times in just over 8 years. - Brian Daniel Eisenberg

“software is lettuce, not gold” (Brian Behlendorf) so perhaps it's never good.. particularly if you are marketing guy facing ever-changing competition - Travis Parsons

I'd say 2-3 years easily, just to get real traction and work out the kinks before the product really starts to mature (if it's any good to start with). If you're out too early and it stinks and everyone bails on it, what's the point? (Case study: Friendster.) - Brandon Uttley

somewhere between 3 days and 3 years - grag

Varies. You can nail it first time (rare) or you can work toward it from a decent base. The latter generally happens in the 2-3 year range. If it takes longer nobody adopts it unless they have to (or if they can change it). - xero

it depends on the company ( big, small or a startup) usually when big/established companies trie to build a product in a market which already has competitors, they try to make it game-changing (a.l.a iphone), or differntiate it ( gmail ) and this means spending sometime on the product to get it right and better. Also, big companies have the advantage of scale means they will have a user base that's willing to try their product out... This could also be a disadvantage, as they would need to worry about all the issues beforehand. That's the reason Google worries about scalability, Microsoft spends months to years testing their operating systems... I guess its different with startups.. how long did it take to build a youtube or a flickr or even a friendfeed ? Months ? Most of the consumer software startups work on their product for a few months and then put it out and iterate over it. - Krishna Gade

Until it's ready. Sort of like toasting a piece of bread. - Ryan Massie

Seems to me like it should be possible to develop software of low complexity in less than the 2-3 year range cited here. For example, Reddit was developed pretty quickly (3 months?), and Twitter must have taken only a few weeks. Of course, once that's successfull, you need to iterate and scale, but that's another topic I think. - Gabor Cselle

Reddit's initial version supposedly took 3 weeks, according to Spez. The fastest I've done anything that got any sort of attention was 1 week - Scrutiny, ArcLite, and Randomicity all took almost exactly that long, and it looks like my FriendFeed interview question will too. Of course, they needed (and sometimes didn't get) lots more time before they'd actually be useful. - Jonathan Tang