A vocal #NoEstimates poster on twitter has a wonderful statement
#NoEstimates is about making decisions and maximizing ROI (Return on Investment)
Except of course for one slight problem
So if we don't know the cost of the investment - the development cost of the software to some level of confidence - then we get a Divide by ZERO error in our ROI formula.
If we have to wait until we've spent all the money to find out what our ROI is, then we have an open ended moving target which is usually called keeping working until the money runs out.
In the end the #NE notion has either been very poorly explained - time boxed scheduling for an assigned fixed budget, short term delivery of value, prioritized by the customer and measures of the performance of the time producing that value to assure we'll get something of value - is a common approach to rapidly emerging requirements. We see this in INTEL system, ERP rapid deployments, rapid stand ups of CRM and the like.
Or the notion is a complete pile of [email protected]#$ and the advocates don't have a clue where, why, or how they'd find someone to buy into this.
I'll assume its the former, and they just haven't found the words yet to convey what they fervently hold to be applicable in their domain will be applicable to others.
Since there has yet to be a case study described in enough detail to assess it's general applicability, it's hard to tell what domain, context, and actual ROI has been produced. This sounds very familiar, like the first days of XP, where just try it was used. That went for awhile, until it ran into the reality of spending other peoples money. In the end it worked out, but only when the XP'ers came in contact with people asking hard questions.
I was in the room when Ron J. was speaking to the US West development team. He discovered just try it needed more explanation than expected. It worked all out and "agile" is now the norm in that OSS/CRM domain. But man was it painful, when the advocates confuse serious questioning with rejection.
