don’t go half-way
via The Seductive Danger Of Half Measures, by Aaron Harris on TechCrunch:
Keeping a conscious eye on what the point of a test or iteration is, not just to itself, but to your overall plan and mission how building a certain number of tutors in a given area influences student activity and community creation, in my case, rather than just the number of tutors removes the halfsies quality of a test. Rather than continually shifting a business strategy to reflect the results of a single test, aggregating data across a set of them, and altering your strategy accordingly creates consistent momentum for your company where the success or failure are equally useful.
Within that framework, there needs to be set decision points — moments where you predetermine that, based on given sets of data, you will make a decision.