To reduce the chances of going too far down the wrong road, we qualify our projects with some sort of business analysis, for example internal rate of return or return on investment or some other fiduciary measurement.  If we are working from a staged-gate project management system, we will relentlessly review our project condition against this defined measure.  Using the staged-gate technique, we suspend our project momentarily and assess the state. We look backwards answering questions about the work we have performed.  Questions such as:

1. Did we answer the questions that this phase of the project was supposed to answer?
2. Did we meet our quality targets for this phase?
3. Did we deliver any models or prototypes on schedule?
   1. Did we learn from those prototypes?
   2. Can we predict future success?
   3. Are we meeting our budget?

We then look forward, striving to answer questions for the next phase of the project. We use questions such as:

1. Can we see success from here?
2. Is there something we should change based upon the last phase?
3. Is this still a worthwhile endeavor?
4. Do we have sufficient resources to continue?
5. Are we seeing interference from other projects?

Being able to derive decisions from the backward look and project valuation from the forward look is one way to help reduce unwelcome surprises.