It seems counter-intuitive to want to fail fast, but if you stop to think about it, it’s far preferable to failing slowly. Killing a bad project early is superior to letting it run and not meet requirements. The faster you fail the sooner you can test another idea, and that may be the one that leads to success, and we know that psychological traps such as anchoring means that you’re likely to keep pushing a bad idea for too long, even if it’s not working out.
What this means:
- Seek feedback early (ideally critical feedback from a potential customer)
- Follow the path that will get you something testable as soon as possible (Frank Gehry builds models out of paper first)
- Determine what your criteria for success are (otherwise you might be tempted to bend the goals if you aren’t hitting your targets)
It seems counter-intuitive to make failure the goal, but getting through a large number of ideas quickly is a better way to find a good idea than plugging along with something that doesn’t work. The overall result will be more successes, faster.
Strategic PPM 
This is precisely the reason why one should deliver value iteratively and in small increments. One can thus get feedback faster and discover the criteria for success sooner. This is one of the main tenets of Agile and Lean Development. Alan Shalloway from NetObjectives prefers the term “learn fast” over “fail fast” – see for example his Blog post http://www.netobjectives.com/blogs/maximum-sprint-length-thirty-days
I think you’re describing the “sunk cost” effect and “loss aversion” more so than anchoring (which primarily impacts estimation, less decision making). From the Wikipedia page on Sunk_Costs: “Evidence from behavioral economics suggests this theory [that a decision-maker may make rational decisions according to their own incentives] fails to predict real-world behavior. Sunk costs greatly affect actors’ decisions, because many humans are loss-averse and thus normally act irrationally when making economic decisions.”
Thanks for sharing Alan’s post, it’s a good idea to re-frame it more positively as learning rather than failing. Agile and lean certainly overlap with this a lot as does the Lean Startup movement, Certainly loss aversion and sunk costs play into people’s reluctance to stop an existing effort, but anchoring is important too in that we are unlikely to adjust perception of benefits that the project will generate (or do it too slowly if we do)..