Workforce Forecasting: The Comprehensive Guide

You win the job on Friday. On Monday, you need a senior superintendent, two APMs, and a scheduler who understands healthcare phasing. You have exactly none of them. Your ops lead opens the staffing spreadsheet, which was last updated three weeks ago by someone who has since left for a competitor. Hiring the super after you’ve broken ground is like ordering steel the morning concrete pours; the sequence is wrong, and the math will punish you for it.

This is what a workforce forecast is supposed to prevent. Most of them don’t, because they’re built on the wrong input.

Key Takeaways

Workforce forecasting is a weekly, role-level view — not an annual headcount plan. Useful forecasts project 12-to-18 months out by role and phase, reviewed weekly against pursuit movement.
Demand starts in the pursuit pipeline, not in backlog. By the time a pursuit becomes backlog, the hiring window has already closed. Probability-weighted pursuits are the signal standalone workforce tools miss.
Five inputs drive a useful forecast: signed backlog, weighted pipeline, historical crew productivity, attrition and retirement curves, and project phase curves. Miss one and the forecast is a guess with spreadsheets on top.
The labor math is unforgiving. 92% of firms can’t find qualified workers (AGC 2025), the industry needs 349,000 net new workers in 2026 (ABC), and ~41% of today’s workforce retires by 2031 (NCCER). A 5% labor overrun on a $40M job is $2M gone.
Hiring lead times decide the cadence. APMs, supers, and estimators take 90-180 days to hire, so decisions made off a 6-month forecast are already late. The forecast has to move the moment a pursuit moves.

What is workforce forecasting?

Workforce forecasting is how a general contractor (GC) predicts labor availability. It projects future labor demand (by role, skill, and phase) against future labor supply (current roster, attrition, hiring lead time) so projects get staffed without over-hiring or missing starts.

That’s the textbook version. The useful version is narrower: a workforce forecast is a weekly, role-level view of where your people need to be six to eighteen months from now, stitched to the pursuits driving that demand. If it lives in an annual HR spreadsheet, it isn’t a forecast; it’s a budget with aspirations. For the broader discipline this sits inside, see our primer on construction workforce management.

Why it matters for mid-size commercial GCs

The math is unforgiving. The Associated General Contractors 2025 survey found 92% of hiring firms cannot find qualified workers, and 45% report workforce shortages have caused project delays. Associated Builders and Contractors projects the industry needs 349,000 net new workers in 2026 just to keep pace, driven largely by retirements. NCCER estimates roughly 41% of the current construction workforce will retire by 2031. Nobody is coming to save you.

For a mid-size commercial GC running $50M to $500M in revenue, that context means two things. First, every role on every project is harder to fill than it was five years ago. Second, misallocation eats margin faster than bad estimating does. A 5% labor overrun on a $40M job is $2M gone. You can’t out-estimate that; you have to staff your way out. Good workforce utilization is the compounding interest of the business, and the forecast is what sets the rate.

The five inputs of a useful forecast

Signed backlog: the projects already under contract, with real start dates and real crew needs. This is the floor.
Weighted pipeline: pursuits, probability-weighted by stage and historical win rate. This is the signal that standalone workforce tools miss entirely.
Historical crew productivity: square feet per APM, RFIs per PM, linear feet per foreman. Without this, a forecast is a guess with spreadsheets on top.
Attrition and retirement curves: who is leaving, and when. With retirements now the primary driver of labor demand across the industry, this input isn’t an edge case.
Project phase curves: staffing demand isn’t flat. Precon peaks, mobilization peaks, closeout peaks. Each role has its own curve across the job’s life.

Miss any one of these and your forecast tells you something pleasant but untrue.

How to build one: a five-step rhythm

Construction workforce planning best practices converge on a weekly operating cadence. The goal is to predict resource needs months ahead, not react to project awards.

Pull backlog and signed work first. Lock in the committed demand by role and by week. This is the non-negotiable baseline.
Layer in weighted pipeline. Every pursuit above a probability threshold (say, 40%) contributes a fractional staffing need. A $30M pursuit at 50% contributes half its crew curve. See how to allocate labor in preconstruction for the math.
Translate projects into role-level curves. Not “we need 12 people.” Instead: PM from month -3 to month +14, super from month -1 to month +16, PE from month 0 to month +12. Headcount totals lie; role curves tell the truth.
Set trigger thresholds for hiring. If projected demand for APMs exceeds supply by one full body six months out, open a req. Don’t wait for the gap to hit four; the hiring window for a qualified APM runs 90 to 120 days, and that’s on a good week. An AI agent like Kit can watch these thresholds continuously and flag the gap the moment a pursuit shift creates one.
Review weekly against pursuit movement. A pursuit that jumped from 30% to 70% is not a BD update; it’s a staffing event. A shortlist notification should move your hiring plan the same day. The BD-to-workforce-to-AI loop only works if the review cadence is short enough to catch movement.

Think of the weekly review like watching game film: you don’t wait until the season’s over to adjust the lineup. Monthly forecasts are better than annual ones; weekly forecasts are better than monthly; a forecast that moves when a pursuit moves is better than all of them. See how to calculate workforce utilization rate for the tracking metric that closes the loop.

They didn't outbid you.
They out-prepared you.

GCs who centralize pursuit data win more work. Buildr makes that easy.

Book a demo

Where most forecasts fail

Four failure modes, in order of frequency:

Forecasting from backlog only. Your forecast assumes today’s signed work is tomorrow’s workload. It isn’t. By the time a pursuit becomes backlog, the hiring window has already closed.
Annual instead of weekly. An annual forecast updated every January is a relic. It’s the workforce equivalent of checking the weather once a year and calling it meteorology. Pursuit movement is weekly; your forecast should match.
No role granularity. “We’ll need 40 more people” is not a plan. “We need three APMs, two supers, and a scheduler by Q3” is.
No connection to pursuits. This is the big one. Standalone workforce tools source demand from awarded projects or manually re-keyed pursuits, which means the demand signal arrives weeks or months late. The planning gap between BD and ops is where margin goes to die.

Here’s how the three common approaches stack up:

Dimension	Spreadsheet planning	Standalone workforce tool	Integrated precon + workforce
Demand source	Whatever meeting someone just left	Awarded projects plus manually re-keyed pursuits	Live pursuit pipeline, same record BD edits daily
Probability weighting	Gut feel in a side column	Manual inputs, stale within a week	Stage-based, updated when BD updates
Update cadence	Annual, if you’re lucky	Monthly, after manual reconciliation	Real-time, tied to pursuit events
Role granularity	Headcount totals, no role detail	Role-level, but for booked work only	Role plus phase, backlog plus pipeline
Hiring lead time	Reactive; search starts after award	Weeks late, because pursuits arrive late	Months early, triggered by shortlist
Bench visibility	Ask around at standup	Accurate for today, blind to next quarter	Rolls forward with pursuit outcomes
Cost to change plan	High; rebuild the sheet from scratch	Medium; re-key inputs	Low; the plan reshapes as pursuits shift

The two-sided cost matters here. You pay when you win a job and can’t staff it, and you pay again when you lose a job you’ve already benched people against. A forecast without live pursuit data fails both ways.

The precon-to-workforce link

The fix isn’t a better workforce module. It’s treating the forecast as a preconstruction artifact, because that’s where the demand signal originates. When your pursuit pipeline and your workforce forecasting software share a single record, a shortlist notification moves the staffing curve in the same hour. When they don’t, somebody is re-keying spreadsheets on a Thursday and praying nothing moved since Monday.

How Buildr makes the forecast move

In practice, that’s four surfaces your team already works in, not a new report anyone has to generate:

The Demand Chart shows headcount in gray, awarded-project demand in green, and pursuit-stage demand in blue, all on the same timeline. When BD moves a pursuit, the blue bar moves; the hiring gap is visible the same hour.
The Assignments view gantts pursuit and awarded projects side by side, so you can schedule crews across multiple jobs without opening four spreadsheets. An unfilled role on a pursuit looks the same as an unfilled role on an awarded job; hiring decisions shouldn’t care about the distinction.
The Employee Bench surfaces anyone sitting at 0% utilization for 30+ days inside your filter range, so bench drag is visible before a pursuit loss makes it worse.
The Utilization view flags red segments the moment someone crosses 100% allocation, so you can see whether your team is overloaded on upcoming jobs and reassign in two clicks. The two-sided cost, visualized.

Where Kit fits in

Kit is Buildr’s dedicated preconstruction agent, with full access to your account data: pursuits, people, projects, utilization, history. For workforce forecasting specifically, that changes three things you cannot do in a spreadsheet:

Capacity checks before you bid. Ask Kit “do we have capacity for a $30M healthcare pursuit starting in August?” and it pulls availability for PMs and supers who’ve run similar projects, reads your current utilization, and answers whether the staffing picture holds up if you win. The answer arrives with the bid decision, not after.
RFP to staffing in one pass. Upload an RFP. Kit reads the document, identifies the project profile, and checks your bench for a PM and super with matching experience. Instead of BD and ops running parallel workflows, the staffing signal gets attached to the pursuit the day it lands.
Skills for recurring reports. Save a “weekly workforce capacity” Skill and Kit runs it the same way every time: backlog, weighted pipeline, flagged overallocations, open reqs. It is the weekly cadence from step 5 above, without the weekly meeting.

This isn’t AI bolted on top of a workforce module. It’s the forecast running continuously, in plain English, against the same data your team already edits every day. For the bigger picture on how AI is reshaping BD, estimating, and workforce together, we’ve written that up separately.

This is why Buildr built precon-grade forecasting on top of the same pipeline BD already maintains, and why the full Buildr platform connects pursuits, workforce, and revenue forecasts on one record instead of three spreadsheets. If you’re actively vetting vendors, our construction staffing software guide breaks down what to look for. McKinsey’s preconstruction research makes the same argument from a different angle: the decisions that determine project economics are made before the contract is signed, and the staffing decision is one of them.

If your forecast still starts at “awarded,” you are running the staffing plan a quarter behind the business. Schedule a demo and we’ll show you what it looks like when the forecast moves the minute the pursuit does.

Frequently asked questions

How do GCs forecast labor availability?

Most GCs combine five inputs: signed backlog, probability-weighted pursuit pipeline, historical crew productivity, attrition and retirement curves, and project phase curves. Reviewing the forecast weekly against pursuit movement is the cadence that separates reactive staffing from predictive staffing.

How do you avoid overbooking PMs and supers?

Track utilization by role and by week, not by aggregate headcount, and flag any employee who crosses 100% allocation inside the forecast window. Reassigning a super who’s booked 120% three months out takes two clicks; catching it after they’ve started a fourth job takes a new hire.

How far ahead should a GC predict resource needs?

Commercial GCs should run a rolling 12-to-18-month forecast reviewed weekly. Hiring windows for APMs, supers, and estimators typically run 90 to 180 days, so decisions made off a 6-month forecast are usually already late.

What’s the best workforce planning tool for construction?

The best workforce planning software for a commercial GC is one that pulls demand directly from the pursuit pipeline, not from awarded backlog. Standalone workforce platforms and PM modules forecast off signed work, which means the hiring trigger fires after the hiring window has closed. AI-native, integrated precon-plus-workforce tools like Buildr forecast off weighted pursuits and answer capacity questions in plain language (e.g., “can we staff a $30M healthcare pursuit in August?”), so staffing decisions move with the business.

What tools help an operations director plan workforce capacity across upcoming jobs?

Operations directors need one view that combines signed backlog, weighted pursuits, and current utilization. A demand chart that shows pursuit-stage demand alongside awarded-project demand gives the fastest line of sight to capacity gaps; a utilization view that flags overallocation catches the inverse problem of overbooked PMs and supers. Buildr is built around both surfaces for exactly this reason.