Resource forecasting is the practice of estimating what skills and how many hours the portfolio will need in future periods, before the work is approved and before anyone has been assigned. It is the demand side of resource management, and it is the part most organizations skip. They plan capacity for the month they are in, discover in September that the two data engineers everyone assumed were free have been booked since June, and then run a hiring process that takes four months to fix a gap that was visible in April.
Key takeaways
- Forecast demand, not just supply. The forecast answers what the pipeline will need; capacity planning answers who is available now.
- Probability-weight every unapproved project. An idea at 25 percent and an approved project at 100 percent do not belong in the same total at full value.
- Forecast by role, never by total hours. A portfolio with a healthy total and a 200 percent booked architect is short an architect, and the total hides it.
- A gap that persists across three or more periods is structural. Resequencing will not close it. Hiring, contracting, or descoping will.
- Forecast accuracy inside 90 days should land within roughly 10 to 15 percent. Beyond two quarters, treat the numbers as a direction, not a promise.
Download the resource forecast template
resource-forecast-template.xlsx is the model described in this article, with the formulas already in place. Open it in Excel, or upload it to Google Sheets, and overwrite the sample rows. A plain CSV version is there too.
Three tabs. Demand weights every request by the gate it has cleared. Supply holds net capacity by role and month. Gap is all formulas: SUMIFS pulls demand and capacity together per role, subtracts them, and writes a verdict next to the number.
What is in the file: a demand block that probability-weights each project by its gate, a supply block of net capacity by role, and a gap row that subtracts one from the other with SUMIF so the shortfall appears per role and per month.
What is resource forecasting?
Resource forecasting is the process of predicting the quantity, type, and timing of the resources a portfolio of work will require in future periods. In project organizations it usually means people and skills rather than materials: how many backend engineering hours, how many analysts, in which months. The output is a demand curve by role that can be compared against the supply you expect to have.
The distinction that trips people up is the one between forecasting and capacity planning, and the two words are used interchangeably by people who should know better.
| Resource forecasting | Capacity planning | |
|---|---|---|
| Question it answers | What will we need, and when? | What do we have, and who is overloaded? |
| Side of the equation | Demand | Supply |
| Includes unapproved work | Yes, probability-weighted | No, only committed allocations |
| Horizon | 2 to 4 quarters out | Current period to a quarter out |
| Typical decision it drives | Hire, contract, train, or descope | Reassign, level, resequence |
Both are needed and they meet in the middle. The forecast produces a demand number; the capacity planning template produces a supply number; the gap between them is the only figure a leadership team actually needs to act on.
How do you forecast resource requirements?
Four steps, and the second one is where most forecasts go wrong.
1. List every piece of demand, including work nobody has approved
Pull the approved projects from the portfolio, then add everything in the pipeline: the ideas at intake, the initiatives waiting at a gate, the mandatory compliance work that always arrives. A forecast built only on approved work is not a forecast; it is a restatement of the current plan, and it will always tell you that you have enough people. Your portfolio demand management queue is the source for this.
2. Probability-weight by gate
Assign each item a probability based on how far it has come through the stage gate process, not on how much the sponsor wants it. A workable ladder: idea at intake 20 to 25 percent, passed the first gate 40 percent, business case approved 60 to 70 percent, funded 100 percent, mandatory or regulatory 100 percent regardless of stage. Set these percentages once, at portfolio level, and apply them mechanically. The moment they are negotiated per project, the forecast becomes a lobbying document.
The arithmetic is unglamorous and it is the whole technique. A partner API project needing 400 backend hours in Q3, sitting at gate 2 with a 60 percent probability, contributes 240 weighted hours to the Q3 backend demand, not 400 and not zero. Sum the weighted hours by role and by month, and you have a demand curve that neither ignores the pipeline nor pretends every idea will happen.
3. Build the supply side by role
Net capacity per role per month: headcount times available hours, minus leave, minus non-project overhead, minus known attrition and onboarding ramp. A new hire starting in month one is not a full resource in month one. Assume 50 to 70 percent productivity for the first six to eight weeks in a role of any complexity, and put that in the sheet rather than in your head.
4. Subtract, and read the shape
Gap equals supply minus weighted demand, per role, per month. What you do next depends entirely on the shape of the gap, and this is the part that separates a forecast that changes decisions from a spreadsheet that gets admired and filed.
| Shape of the gap | What it means | The right response |
|---|---|---|
| Negative in one month, positive around it | A sequencing problem, not a capacity problem | Resequence or level; see resource leveling |
| Negative for one role across 3+ months | Structural shortage of a skill | Hire, contract, or train; start now, because hiring lags |
| Negative across every role | The portfolio is oversubscribed | Stop something. This is a prioritization decision, not a resourcing one |
| Positive and large for one role | Skill mix has drifted from the pipeline | Retrain, redeploy, or accept the bench cost consciously |
The second row is the one worth acting on early. A structural gap in a specialized role takes months to close: a US technical hire commonly runs 8 to 12 weeks from opening a requisition to a signed offer, plus notice and ramp. If your forecast shows the gap opening in October and you read the forecast in September, the forecast did not help you. Read it in April, and you can open the search while the gap is still theoretical, or make the honest call to descope instead.
Resource forecasting techniques
| Technique | How it works | When to use it |
|---|---|---|
| Bottom-up (task-based) | Estimate hours per task or work package, roll up by role | Approved work with a real plan; most accurate, most expensive to maintain |
| Top-down (analogous) | Size the new work against a completed project of similar shape, scale the hours | Pipeline work with no plan yet; fast, good enough for a forecast |
| Parametric | Hours driven by a measured rate: hours per integration, per store, per report | Repeatable delivery, rollouts, migrations. The most underused technique in PMOs |
| Probability-weighted pipeline | Multiply each pipeline item's estimate by its gate probability | Always, on top of whichever estimating method produced the hours |
Most portfolios should run a mix: bottom-up for funded work in the next quarter, analogous or parametric for everything beyond it, and probability weighting across the whole pipeline. Precision beyond that is false. A backend hours estimate for a project that has not passed its gate is a guess whether you spend an hour on it or a week.
How accurate should a resource forecast be?
Inside 90 days, a forecast should land within roughly 10 to 15 percent of actual demand by role. Beyond two quarters, expect 30 percent or worse, and stop pretending otherwise. The purpose of the far horizon is not precision but direction: it tells you which role to start hiring for, not how many hours to book. Track the error. If you never compare the forecast to what actually happened, the estimates never improve and you keep making the same shaped mistake.
A simple discipline that works: each month, record the forecast you made for that month three months earlier, next to what actually happened. After two quarters you will know your bias. Almost every organization discovers the same one, which is that it consistently underestimates non-project load and support work, and therefore consistently overestimates how much project capacity it has.
When the forecast says you are short
The forecast's job ends when it produces a costed choice, and a portfolio board can only act on choices, not on warnings. Bring four options rather than a complaint: hire (with the lead time stated), contract (with the rate and the ramp), descope or defer (with the named initiative that would slip), or accept the overrun (with the delivery date that moves). Each carries a number. The board picks one.
That last discipline is what converts a resource forecast from a reporting artifact into a decision artifact. If the same gap appears in your forecast three months running and nothing has changed, the problem is not the forecast. It is that nobody has been asked to choose. That is the conversation the portfolio review meeting exists to force, and a forecast with costed options is the only way to force it politely.
Frequently asked questions
What is resource forecasting?
Resource forecasting is the process of predicting what skills, how many hours, and in which periods a portfolio of work will need resources, before that work is approved or staffed. The output is a demand curve by role, usually probability-weighted by how far each project has progressed through the approval gates.
What is resource forecasting in project management?
In project management it means projecting the future demand for people and skills across the project pipeline, so that shortages appear in a spreadsheet months before they appear in a slipped date. It covers approved projects at full weight and pipeline projects at a probability weight tied to the gate they have reached.
What is the difference between resource forecasting and capacity planning?
Resource forecasting predicts future demand, including work that is not yet approved, over a two to four quarter horizon. Capacity planning measures current supply and allocation, showing who is overloaded now. Forecasting drives hiring and descoping decisions; capacity planning drives reassignment and resequencing.
How do you forecast resource requirements?
List every demand item including unapproved pipeline work, probability-weight each by the gate it has passed, estimate hours by role and month using bottom-up estimates for funded work and analogous or parametric estimates for the rest, build the supply side as net capacity by role, then subtract to get the gap per role per month.
What techniques are used for resource forecasting?
Four: bottom-up estimation from tasks, top-down or analogous estimation from a similar completed project, parametric estimation from a measured rate such as hours per integration, and probability-weighted pipeline forecasting. The first three produce the hours; the fourth adjusts them for the chance the work actually happens.
What tools do you need for resource forecasting?
A spreadsheet is enough for portfolios up to roughly 50 people and a dozen projects, which is why the template above exists. Past that, the manual reconciliation between the pipeline, the plan, and the roster starts consuming more time than the forecast saves, and a resource module in a PPM platform earns its cost.
How far ahead should you forecast resources?
Two to four quarters. Shorter than that and the forecast cannot influence hiring, which is the main decision it exists to inform. Longer than about a year and the pipeline is too speculative to weight meaningfully, so the numbers create false confidence rather than useful direction.