Scheduling Engine — Who Goes First?
Scheduling Engine — Who Goes First?
You already know the demand list. You already know who's on the team. What you don't have is a defensible answer to the two questions that decide your week:
Who goes first when two demands want the same person — and what do you tell the one who loses?
The 30-second version
*Same inputs, two outcomes. Top: the manual pass loads nine days of work onto Alice and still expects a Friday finish — the Animation work simply never lands. Bottom: the engine respects her capacity, keeps it on the chart, and tells you the real date. Example run.*
Nine days of work, three people, one Friday deadline.
A hand-built plan treats people as if they were machines with no capacity limit. The engine doesn't — it reads the workload per person, keeps every piece of work visible, and returns the date you will actually hit, plus the name of the person who is the constraint.
Three questions this answers
1 · Sequencing — who goes first?
Example run — per-resource load with the bottleneck verdict. The real constraint is the serial skill chain, not just the busiest-looking name.
When two demands claim the same engineer or the same slot, you get a decision with a reason you can check and override — not a silent reorder.
2 · Conflict arbitration — what do I tell the person who loses?
Every contested resource produces an explicit verdict: who waits, how long, and on what grounds. That is the sentence you actually have to say out loud in the room.
3 · Deadline risk — will Friday hold?
Example run — the CPM / float table. The longest task is not automatically the critical one; the chain with zero float decides the ship date.
You learn that a date is impossible before you promise it. The engine states it plainly, with the number attached:
Baseline: makespan 9 days against a 5-day deadline — 4 working days late.
Bottleneck: Alice (backend) — her serial chain occupies all 9 days.
Overrun item: R5 Animation — 4 days past the deadline, flagged as "pending weekly re-plan", never silently dropped.You get the honest number, the responsible name, and the work that has to move. The trade-off decision stays yours.
What you get
- 📊 A dated Gantt schedule — real calendar dates from your real availability, not a flat task list
- ⚖️ Resource conflicts decided — who goes first, who waits, and the reasoning behind it
- ⚠️ Deadline slippage caught early — flagged before you commit the plan to anyone
- 🧭 The zero-float critical path — the chain that actually decides the date
- 🔎 Bottleneck attribution with a name — not a load percentage, the person carrying the critical chain
- 🧾 A per-person workload that shows the overrun — so "we're not going to make it" comes with evidence
- 📐 Estimates for unestimated work — from historical analogues, with an honest confidence label
📖 Reading the float column
The Gantt shows total float (TF) per task, computed on a forward/backward pass that includes both dependency arcs and same-person resource arcs.
Why resource arcs matter: when one person owns two tasks, the second cannot start until that person is free — even with no dependency between them. Omit that arc and waiting-for-a-person time is misread as float, and you get told there is slack that does not exist.
TF = 0 means: this task is on the critical path. If it slips, the deadline slips.
TF > 0 means the project still has slack — but float can sit on a busy person's chain, so read the bars for the real hand-off order, not the float column alone.
⚠️ What it warns you about
| Warning | What it means for you |
|---|---|
| Deadline slippage | You learn who loses before you promise them anything |
| Bottleneck single point | One sick day = the whole version slips a day |
| Idle-gap defect | A short schedule with a broken-looking chart is still a broken schedule |
| False-conflict alarm | Stops you from "fixing" a schedule that was already correct |
| Optimism trap | Ships the date you can hit, not the one that looks best |
Who it's for
Anyone who has to sequence shared resources against a deadline:
- Project / production PMs and PMOs sequencing a version, a milestone, or a release train
- Team leads and tech leads answering "can we take this on, and what slips if we do?"
- Studio owners / small teams where one person covers three roles and every conflict is personal
- Producers on non-game projects — the method is about contested resources, not about games
Run your first project free in under a minute.
The question it answers is always the same: who wins the contested resource, and what do I tell the person who loses?
