There is a pattern every systems engineer recognizes. At SRR, the mass budget shows 30% margin. By PDR, it is 20%. By CDR, you are fighting for 10% and reclassifying items to make the numbers work. The spacecraft did not get heavier overnight. The estimates got more honest.
The problem is how mass budgets are typically managed. In a spreadsheet, each subsystem lead enters their best estimate. They round down because no one wants to be the reason the design does not close. Harness mass gets estimated at 5% of dry mass when the actual number is closer to 8%. Thermal blankets get a placeholder that never gets updated. The margin looks healthy because the estimates are systematically biased low.
A connected mass budget fixes this by tying allocations to actual component selections, not estimates. When you select a specific reaction wheel, the mass budget pulls the datasheet mass, not an engineer's guess. When the harness routing is defined, the budget reflects measured lengths and wire gauges. The margin is real because the inputs are real.
SMAD Portal's mass budget tracks allocations at the component level with margins computed at each tier. When a subsystem allocation changes, the total margin updates immediately. You see the design close or not close in real time, not at the next monthly program review.