In my previous post, Better products faster, a reaction to this discussion on linkedin, my main point was: how do you define success for Agile?
But before you can define success, you need to specify what Agile is. Even this seems unclear. One reply on the discussion calls it a marketing/branding term in the IT sector. Also common sense, pragmatism, and an adaptive approach to project management. For others doing agile means following the principles of the Agile manifesto.
I see Agile mostly as a reaction to the failures of waterfall: trying to predict long running, complex projects. Following a predefined plan, not being able to alter requirements or plans when new information or situations arise.
Waterfall is behaving like you have a crystal ball, Agile is the trial and error way of reaching your goals.
Calling it trial and error is probably not the best way to sell it. PDCA (plan, do, check, act) or empiricism are better names from a marketing point of view.
It all comes down to not pretending that you know the future. Instead you take small steps, determine the results of these steps, and you adjust your plan according to the feedback you collected.
Donald Reinertsen makes an interesting statement about trial and error in his book Managing the design factory: The try-it-fix-it approach is faster and higher quality (page 76).
This is agility: being able to adjust your plans based on new information.
Not just from the perspective of the developers, but even more so for Product Managers.