Killing projects needs an innovative company culture that accept mistakes and embrace challenges critically (no bias).
Sunk costs should not be part of the go-no go decision
A regular assessment of the project end-point together with the routine progress tracking could be beneficial.

Knowing when to stop a project is a key element of good project management and this article is a good refresher on this topic.

I would like to refere to the following part of the article:

"People's natural inclination is to focus on the positive," Shimamoto said. "The question is 'How do we save this project?' instead of 'Should we be killing this project?' "

I would say that it is not bad to ask at first how we could save the project. There is always some solution to the problem. Of course, sometimes one of these solutions my be just to kill the project. However, at first, I would check all other posibilities how project could be improved or modified so that at the end of the story there is some benefit out of this project.

Very seldom is there a culture of introspection and reflection within organizations, particularly larger organizations to determine if a project was a success. I think if there was more of a review I think the organizations would be more capable of stopping projects that don't add value. Generally whether a project should be stopped should come down to value. For what we are paying (cost), for the quality (scope) we are getting and for the time it will take is it adding value to our business. 

Value does not necessarily have to equate to monetary value but you can't do many projects that don't add to your financial bottom line. 

Killing or stopping a project should be an integral part of Project Management. Knowing when to stop a project, and to take the decision and action to stop it is critical to avoid losing and wasting more resources than necessary.

Decisions to kill a project a usually delayed by softer issues like ego's  and feelings of the people involved.

To ensure that this decision is not delayed, there should be a role within each project whose purpose is to ask the question 'should we continue or kill the project', based on current situation.

There are many examples of projects that took long to be stopped, and did not realize the envisaged return. Killing a project when required is not failure, but failing to kill a project when all indications are that it should be- is failure.