A master schedule tells you what should happen. Short interval planning tells you what can actually happen this week, given who showed up and what actually arrived — and the two documents disagree more often than most schedules admit.
Key takeaways
Lean Construction Institute – Canada was established on June 5, 2015 as a special committee of the Canadian Construction Association, with a stated purpose of applying lean tools and techniques throughout the building industry supply chain to improve the value delivered by capital projects while reducing waste. Short interval planning is one of the practical mechanics that lean methodology runs on.
The clearest description of the mechanics comes from the Last Planner System: a lookahead plan identifies the activities that should be ready to run over the next several weeks, and a weekly work plan turns that into specific, collaborative commitments — weekly promises between trades about what will actually get done, made and tracked at a short interval rather than a monthly one.
A master schedule is built on assumptions made months before the work happens: that materials arrive on time, that the trade ahead of you finishes on schedule, that nobody's short a crew that week. A short interval plan doesn't replace those assumptions — it re-tests them weekly against what's actually true right now.
That re-testing only works if the data behind it is current. The same daily-log categories a site already tracks for other reasons — manpower, delivery, equipment and scheduled-work entries — are exactly what a short interval plan needs as inputs: not a monthly report, but what actually happened yesterday, feeding what's realistic tomorrow. The same daily record that feeds this weekly re-test is also what makes a site diary defensible later — the two disciplines reinforce each other rather than competing for someone's time.
Short interval planning is a weekly, collaborative discipline — it needs someone reading the real conditions and running the commitment conversation with trades, in person or close to it, every week, on every site it covers. That's exactly the capacity problem covered in supervising more sites than you should: a superintendent spread across enough sites to only touch each one occasionally can run a lookahead schedule on paper, but can't run the weekly promise-and-track conversation that actually makes it work.
The practice that separates a real short interval plan from a schedule nobody looks at is the weekly review of last week's promises against what actually happened — not to assign blame, but to feed the next week's plan with what's actually true about capacity, sequencing and delivery reliability on that specific site.
That review works best as a short, standing meeting with the trades whose promises are being checked, not a report circulated after the fact. A promise that slipped because of a late delivery is a different problem — and gets a different fix — than one that slipped because two trades were scheduled into the same physical space at the same time. Only a live conversation with the people who made the promise reliably tells you which one happened.
A working session doesn't need to be long to be useful. At minimum it should walk through: which of last week's commitments were kept, which weren't and why, what's actually ready to start this week given current manpower and material on hand, and what this week's crews are explicitly committing to by Friday. Kept short and specific, that's a brief standing conversation, not a half-day planning exercise — the value is in the weekly repetition, not the length of any one session.
Multi-trade sites add one more thing worth checking every week: whether this week's plan actually has each trade in a different physical space at a different time, or whether two crews are quietly counting on the same square footage. A lookahead built from each trade's own plan in isolation won't catch that; it only shows up when the plans are checked against each other in the same room.
A worked example
A framing crew commits, in Monday's weekly plan, to closing in a section by Friday, contingent on a lumber delivery scheduled for Wednesday. The delivery slips to Thursday afternoon — logged the same day it happened, not reconstructed later.
Because the delay is in the log, Friday's promise gets revised honestly on Thursday rather than discovered as a missed deadline on Monday: the crew shifts to prep work that doesn't depend on the delivery, and the following week's plan absorbs the actual one-day slip instead of pretending it didn't happen. The master schedule doesn't need to move for a one-day input delay — the short interval plan is what keeps that delay from compounding silently into a bigger one.
Compare that with the same delay on a site with no short interval discipline: the delivery slips, nobody re-plans around it because there's no weekly checkpoint to catch it, and the mechanical trade scheduled to follow framing shows up Monday to a section that isn't ready. The one-day input delay has become a one-week schedule problem, and the master schedule is the only document anyone notices has slipped — days after the actual cause.
No. A master schedule sets the long-range sequence and milestones for the whole project; short interval planning re-tests a short window of that sequence — typically the next one to several weeks — against actual, current site conditions. The Last Planner System treats the lookahead plan and the master schedule as related but distinct documents.
A special committee of the Canadian Construction Association, established June 5, 2015, focused on applying lean planning and delivery methods across the Canadian construction industry.
No — the method predates any specific platform. What it needs is a reliable, same-day record of manpower, deliveries and completed work, and a weekly forum for trades to commit to and review specific promises. Software makes the log faster to keep current; it isn't what makes the method work.
Long enough to cover last week's kept and missed promises, this week's ready work, and this week's new commitments, and no longer than that — a short, standing conversation run every week without fail, not a lengthy session reconstructed from scratch each time. The value comes from the weekly repetition, not from making any single session longer.
Capacity, more often than resistance to the method itself. Running a real weekly session on every site it covers takes a planner or supervisor who's actually present often enough to hold the conversation — which is exactly where it collides with how many sites one person is realistically covering.
A 30-minute call is enough to tell you whether AI pays for itself here.
