🎯 The Goal: Most tasks don't actually dictate your deadline — a hidden chain of them does. In every project there's one continuous sequence of dependent tasks where ANY delay pushes the final date: the critical path. Everything else has slack. Teams miss deadlines because they treat all 40 tasks as equally urgent, sprint on the easy ones, and starve the chain that actually decides the ship date. Find the chain; guard the chain.
🗺️ Step 1: Map the dependency chain (30 minutes, whiteboard or sticky notes). 1. List every task in the project. 2. For each, ask the one structural question: 'What must be FINISHED before this can START?' Draw the arrows. 3. Estimate each task's duration honestly (use the team's worst-recent-project numbers, not their optimism). 4. Trace every path from start to launch and ADD UP the durations — the LONGEST total path is your critical path, and its sum is your realest possible ship date, no matter what any spreadsheet wishes. Everything off that path has slack (it can slip days without moving launch); everything on it has zero. Most people are shocked twice: by which tasks are actually critical (the permit approval, the client's content delivery, the one integration) and by how many 'urgent' tasks aren't.
🚦 Step 2: Resource the path like it's the whole project — because it is. Operating rules: 1. Your best people work the critical path FIRST every day; slack tasks get the leftover hours (an hour saved on the path = launch moves a day earlier; an hour saved off-path = nothing happens). 2. Critical-path tasks never wait in queues — reviews, approvals, and answers for path tasks jump every line, and everyone knows why. 3. Watch the handoffs hardest: the classic silent killer isn't slow work, it's finished work sitting 3 days waiting for the next person to notice ('done Friday, picked up Wednesday' just ate your buffer). Make every path handoff an active ping, not a passive inbox. 4. External dependencies on the path (client assets, vendor parts, approvals) get started EARLIEST and chased weekly — they're the tasks you control least on the chain that matters most.
🔭 Step 3: Re-check the path weekly — it moves. When a path task finishes early or a slack task slips badly, the longest chain can REROUTE through different tasks entirely; a 10-minute weekly re-trace keeps you guarding the real one instead of last month's. More field rules: 1. Buffer goes at the END of the path (one shared reserve), not padded secretly into every estimate — hidden padding gets consumed invisibly (work expands); a visible end-buffer gets defended by everyone. 2. Shortening the path is the ONLY way to ship earlier: split a path task so halves run parallel, start a downstream task at 80% handoff instead of 100%, or move a path task off-path by killing its dependency. Adding people to off-path work changes nothing but the burn rate. 3. Solo/small projects: same math, humbler tools — even a 10-task freelance gig has a chain ('client content' is nearly always on it; chase it from day one, see the scope lesson's cousin problem).
📋 The Critical-Path Cheat Sheet: 1. Map dependencies → longest chain of durations = your real deadline. 2. Path tasks: best people, first hours, front of every queue. 3. Handoffs are the silent killer — active pings, never inboxes. 4. External path dependencies: start earliest, chase weekly. 5. Re-trace the path every week; it reroutes. 6. One visible buffer at the end; speed comes only from shortening the chain. Not all urgency is real — the chain decides. Find it, feed it, ship on time.