How to set milestone schedules that actually predict ship dates
Most milestone schedules are wishful thinking that gets discovered to be wrong two weeks before ship. Here’s how to set ones that hold up.
By Kuhler Operations Team · Operations
A milestone schedule is a contract with reality. Most apparel ones aren’t — they’re a salesperson’s commitment with a factory’s nodding. Real schedules need to predict ship dates, and most don’t.
Why most milestone schedules fail
- They’re back-cast from the customer’s requested ship date, not forward-cast from material readiness
- Buffers are added at the end (right before ship) instead of distributed across the riskiest steps
- No measurement of past variance per step — same factory, same product, no learning
- Status is binary (on/off track) instead of probabilistic (90% / 70% / 50% likely)
A schedule you can’t miss isn’t a schedule. It’s a wish.
The five-milestone backbone
For most cut-and-sew apparel, five milestones cover 95% of the risk:
- Materials sourced — fabric, trim, accessories all confirmed and en route to factory
- Cut & sew started — production has begun on the first piece
- Production midpoint — half the units off the line, QC inline, no fabric or process issues found
- QC pass — inspection complete, defects below threshold, ready to pack
- Pack & ship — packed by SKU, paperwork done, on the truck
Each milestone has a target date AND a confidence level (high / medium / low). When a milestone slips, the schedule updates automatically — not by moving the ship date, but by re-estimating downstream confidence based on this slip’s pattern.
The 7-3-1 buffer rule
Distribute schedule buffer across the three riskiest steps, not at the end:
- 7 days at materials sourced — international shipping, fabric mills, trim suppliers all have variance
- 3 days at production midpoint — defect discovery, line speed adjustments, second-batch issues
- 1 day at pack & ship — paperwork, customs, last-minute label changes
Total: 11 days of buffer, distributed where slips actually happen. Most schedules put 14 days at the end and discover the slip is unrecoverable on day 13.
Measure the variance
Atlas tracks every milestone date — target, actual, slip days — by factory and by product family. After 10 batches you have a real distribution per step. The next schedule for that factory + product can be set against historical performance, not against the salesperson’s optimism.
After 50 batches, you have factory-specific predictability scores. You can quote ship dates with confidence intervals. You can spot the factory that’s slipping ships before the customer does. That’s the whole game.