Two things quietly happen inside any consulting sprint longer than eight weeks: it scope-creeps, and the last third gets spent doing what should have been done in the first.
Six weeks is not a magic number. It's the point at which two things are simultaneously true: you can ship a real vertical slice of production software, and you cannot pretend the deadline is far away.
What twelve-week engagements actually look like
Weeks 1-3: setup, alignment, discovery. Weeks 4-8: real work. Weeks 9-11: scope creep and re-planning. Week 12: launch scramble. The math never survives contact with reality.
What we do instead
Break work into consecutive six-week sprints, each with a shipped artifact at the end. If you need three sprints, we'll bid three. But the sprint boundary is not administrative, it's a real shippable outcome.
The exception
A SaaS build with dozens of screens can't compress below eight weeks and stay coherent. For those, we run 8-12 weeks, but with a working demo at week 2 and a soft launch at week 6.