When is the right time to "ship" your product?
Prior to just a couple of years ago, the typical software development model was to ship the product when it's done. That typically involved many steps and processes over an extended period of time (less than a year from planning to live product was considered fast.) Activities included:
The big question in an iterative agile development model is WHEN is the right time to "ship" the product to the public? Do you stay stealth until you have the product fully baked -- that is say after a handful of iterative cycles and private alpha user testing? Or, do you go public early and iterate in front of everyone?
We're thinking a lot about these questions here at social|median. After just 4 weeks of software development, we shipped our first "dogfood" code this past Friday and expanded from 10 to about 50 early users. It's a good start but it's not what you'd typically call ready for "prime time." It works though and already provides some instant utility. We know the product isn't even close to being done (heck, we've completed less than 10% of our known early features and we haven't applied any graphic design what-so-ever), but it sure is fun and incredibly helpful to see people touching the code while we're still working on it. So, we're thinking of adding more users very quickly. And, we're thinking of just throwing it out there really early and just iterating it in real time in front of the world.
It's a risky path as some folks are bound to use the early product, not like it, and never come back.
What do you think?
- Plan it
- Prototype it
- Spec it
- Design it
- Build it to spec
- QA it
- User test it
- Revise it
- QA it
- SHIP IT
The big question in an iterative agile development model is WHEN is the right time to "ship" the product to the public? Do you stay stealth until you have the product fully baked -- that is say after a handful of iterative cycles and private alpha user testing? Or, do you go public early and iterate in front of everyone?
We're thinking a lot about these questions here at social|median. After just 4 weeks of software development, we shipped our first "dogfood" code this past Friday and expanded from 10 to about 50 early users. It's a good start but it's not what you'd typically call ready for "prime time." It works though and already provides some instant utility. We know the product isn't even close to being done (heck, we've completed less than 10% of our known early features and we haven't applied any graphic design what-so-ever), but it sure is fun and incredibly helpful to see people touching the code while we're still working on it. So, we're thinking of adding more users very quickly. And, we're thinking of just throwing it out there really early and just iterating it in real time in front of the world.
It's a risky path as some folks are bound to use the early product, not like it, and never come back.
What do you think?
Categories
social|median , startups0 TrackBacks
Listed below are links to blogs that reference this entry: When is the right time to "ship" your product?.
TrackBack URL for this entry: http://blog.socialmedian.com/cgi-sys/cgiwrap/jasong2o/managed-mt/mt-tb.cgi/238
Posted by JD | March 10, 2008 9:39 AM | Score: 0
I believe it really depends on if you have hit a broad enough swatch of your prospective audience. Once you get bug count down it really comes down to being able to exposure the top 10-20% of features that users are most likely to use. I think the testing needs to get out to 100s if not 1000s of users to see the full spectrum of usage. Take a clinical trial for example, pharma companies basically go through 3 phase of testing before rolling out to public. First the anminals then phase 2 the desperate humans (100s) then phase 3 to thousands. By the time it gets to 1000s your really able to refine the software so that you can have the broadest possible adoption.
Posted by JD | March 10, 2008 9:40 AM | Score: 0
I believe it really depends on if you have hit a broad enough swatch of your prospective audience. Once you get bug count down it really comes down to being able to exposure the top 10-20% of features that users are most likely to use. I think the testing needs to get out to 100s if not 1000s of users to see the full spectrum of usage. Take a clinical trial for example, pharma companies basically go through 3 phase of testing before rolling out to public. First the anminals then phase 2 the desperate humans (100s) then phase 3 to thousands. By the time it gets to 1000s your really able to refine the software so that you can have the broadest possible adoption.