Q: Can this project launch in two weeks?
A: Don't ask me to promise off the top of my head. Let's spread the two weeks out and count them — you'll see which parts you can help us skip.
Q: How do you spread it?
A: A table. 14 days.
- Requirements confirmation and prototyping: 2 days. In those 2 days, fields, definitions, and who signs off get locked down.
- Development: 5 days. Against the confirmed scope, excluding "oh, and one more thing".
- Integration testing: 2 days. Aligning with your accounts, data sources, and APIs — the most underestimated stretch.
- Testing: 2 days. One functional pass, one regression pass.
- Acceptance and buffer: 3 days. Addressing feedback, launch, and some slack.
That's exactly 14 days, with no idle day in between.
Q: So can it be done or not?
A: Yes, with three conditions. One: the scope is finalized today — anything added later goes to the next phase. Two: your materials, accounts, and API contacts arrive on the agreed dates; a one-day slip pushes everything downstream by a day. Three: your acceptance people are available those days — don't save it all for the last night.
Of the three, the second fails most often. On a previous project we waited 4 days on a single Excel file — development had actually finished long before.
Q: What if we absolutely must compress it further?
A: Things can be cut, but be clear about what's being cut. We once did rush a project; what got cut in the end was two secondary pages and one export format — not testing, not the buffer. Cut the 2 testing days, and post-launch rework usually runs past 2 days.
Q: Have you ever compressed something and regretted it?
A: Yes. Once we squeezed integration testing from 2 days into half a day; on launch day we found their API returned a different set of fields in production — one extra nesting level compared to the test environment. It took 3 hours to fix. We have never touched that cell since.
Q: So really, you're negotiating terms.
A: That's right. A schedule isn't a promise — it's the word "two weeks" converted into 14 cells, each with an owner and a fallback when it slips. Once it's spread out, both sides know whether it can be done. That 14-cell table later became a template: next time we discuss a deadline, we just fill in numbers, copy the three conditions verbatim, and skip the argument from scratch.
