Project Controls in the AI EraForecasting cost and schedule outcomes · Lesson 9 of 22
EAC and ETC: forecasting the final cost
Video lecture
EAC and ETC: forecasting the final cost
The narrated lecture is in production
Every chapter is scripted and ready. Browse the chapters and read the full transcript now — the video will appear here when it’s published.
Chapters
Transcript of the narration, chapter by chapter.
0:00 A forecast is a judgement, not a formula
Here's a pattern you've probably seen. For eight months, the project forecast equals the budget, exactly. Then in month nine, it jumps by twenty per cent, and everyone acts surprised. That's called hockey-stick forecasting, and it happens when a forecast is treated as a promise rather than an estimate. In this lecture you'll learn the standard formulas for the estimate at completion, what each one assumes, and how to use the to-complete performance index to challenge a forecast that simply isn't credible. Then I'll show you how to build a bottom-up estimate to complete and choose a final number with a rationale. By the end, you'll be able to walk into a forecast review with a defensible range and one clear recommendation.
0:53 Why it matters
Why does forecasting matter so much? Because decisions hang on it. Funding requests, contingency releases, contract negotiations and even whether to continue the project at all depend on the estimate at completion. And timing is everything. Recognise an overrun in month three and you have lots of options: redesign, re-sequence, renegotiate, change the scope. Recognise it in month nine and you have one option: ask for more money. Formulas matter here because they're independent. They don't care how confident the project manager feels. They give you a statistically grounded check that you can hold up against the team's own estimate, and that's often enough to start an honest conversation.
1:40 The concept: four EAC methods
Start with the identity: estimate at completion equals actual cost plus estimate to complete. The question is how you estimate the remaining cost. Here are the four standard approaches, and each comes with an assumption. The budget-rate method: actual cost plus budget at completion minus earned value. It assumes past variance was a one-off and the remaining work will perform to budget. The CPI method: budget at completion divided by CPI. It assumes current cost efficiency continues. The composite method: actual cost plus the remaining work divided by CPI times SPI. It assumes both cost and schedule pressure will affect the remaining cost, for example because you'll need to accelerate. And bottom-up: actual cost plus a fresh, detailed estimate of the remaining work. Think of them like three different weather models plus a local forecaster's view. When they agree, you're confident. When they don't, you ask why.
2:44 Worked example one: quick EAC
Let's start with a simple example. Budget at completion, one hundred thousand. Earned value, forty thousand. Actual cost, fifty thousand. So CPI is forty over fifty, nought point eight. Remaining work at budget rates is one hundred minus forty: sixty thousand. Budget-rate EAC: actual cost fifty plus sixty. One hundred and ten thousand. CPI method: one hundred divided by nought point eight. One hundred and twenty-five thousand. So we already have a range, one ten to one twenty-five. Which is right? That depends on the cause. If the overspend was a one-off event, one ten is plausible. If it's a productivity problem that will continue, one twenty-five is more realistic. The formula can't tell you. The root cause can.
3:36 Worked example two: Karachi Gateway Logistics
Now the realistic example. Karachi Gateway Logistics again: budget two million, and at month four, planned value eight hundred thousand, earned value seven hundred thousand, actual cost eight hundred and fifty thousand. CPI is about nought point eight two, SPI nought point eight seven five. Remaining work: one point three million. Budget rate: eight fifty plus one point three million, two point one five million. CPI method: two million divided by nought point eight two three five, about two point four three million. Composite: eight fifty plus one point three million divided by about nought point seven two, roughly two point six five million. Now the reality check. The to-complete performance index is the remaining work divided by the remaining budget: one point three million over one point one five million. About one point one three. The project has achieved nought point eight two so far. To hit budget, it must suddenly perform at one point one three. Without a fundamental change, that's not a plan. It's a wish.
4:49 TCPI: the reality check
Let's dwell on TCPI, because it's one of the most useful tools in any forecast review. To-complete performance index to budget equals budget minus earned value, divided by budget minus actual cost. In plain English: the value of the remaining work divided by the money that's left. You can also calculate it against an approved or proposed EAC: budget minus earned value, divided by EAC minus actual cost. Then compare it with the CPI you've actually achieved. If a manager forecasts on budget with a CPI of nought point eight four, and the implied TCPI is one point two one, you ask one simple question. What specifically will change to deliver a forty-four per cent improvement in efficiency? If there's no specific answer, the forecast needs to change.
5:45 Watch me do it: forecast sheet in Excel
Let me show you the forecast sheet I use. Inputs at the top: budget, planned value, earned value and actual cost for the control account. Then CPI and SPI. Then three cells for the formula EACs, each with its formula visible so reviewers can check it. Budget rate is AC plus BAC minus EV. CPI method is BAC divided by CPI. Composite is AC plus BAC minus EV, divided by CPI times SPI. Next, the TCPI to budget, formatted red if it's more than about ten per cent above current CPI. Below that, the bottom-up section: remaining labour, materials, subcontracts, pending changes and a risk allowance. Finally, the selected EAC and a two or three sentence rationale. For Karachi, I'd write: selected EAC two point four five million, range two point one five to two point six five. The main swing factor is whether steel productivity recovers when the new crew starts.
6:51 Bottom-up ETC and choosing the number
The bottom-up estimate to complete is where the team's knowledge comes in. List the remaining scope by work package. Re-estimate each one using current productivity, current prices and any known changes. Add allowances for approved changes not yet in the baseline, and pending changes that are likely to be approved. Then add the remaining risk exposure from the risk register or your quantitative analysis. Now compare the resulting EAC with the formula range. If the bottom-up figure lands below the most optimistic formula, that's not automatically wrong, but you need a clear, evidenced explanation, such as a one-off problem that's genuinely been fixed. Experience across many projects suggests cumulative CPI tends to be fairly stable once work is well underway, so claims of dramatic recovery deserve healthy scepticism.
7:47 Common mistakes
The common mistakes. The hockey stick: forecast equals budget until the very end. Using the budget-rate method for a problem that's clearly systemic, such as low productivity. Ignoring pending changes and claims that everyone knows are coming. Presenting a single number to the pound, with false precision, instead of a range with a rationale. And forecasting only at project level, where a healthy civil package hides a failing mechanical one. A quick word on AI: machine-learning models trained on your organisation's past projects can produce an independent EAC and flag when a team's forecast is statistically unusual. Treat that as a third opinion alongside the formulas and the bottom-up estimate, and document which forecast was adopted and why.
8:38 Recap and try this now
Let's recap. Estimate at completion equals actual cost plus estimate to complete. The budget-rate, CPI, composite and bottom-up methods each rest on an assumption, and the root cause of the variance tells you which assumption is realistic. TCPI is your reality check: if the required efficiency is far above what's been achieved, the forecast isn't credible. And when you present, show the range, the selected figure, and the one or two assumptions that would move it most. That invites a useful discussion instead of an argument about a single number. Your try-this-now: using the Karachi example, write a three-sentence forecast rationale. Recommend one EAC, explain why, and explain why you didn't choose the other methods.
A forecast is a professional judgment, not a formula
The estimate at completion (EAC) is the expected total cost when the work is done. The estimate to complete (ETC) is the expected cost of the remaining work, so EAC = AC + ETC. Formulas give you independent, statistically grounded checks; the final forecast is a judgment informed by those checks and by what the team knows about the remaining work.
The main EAC formulas
| Method | Formula | Assumption |
|---|---|---|
| Budget rate (atypical variance) | EAC = AC + (BAC − EV) | Past variance was a one-off; remaining work performs to budget |
| CPI method | EAC = BAC / CPI | Future cost efficiency continues at cumulative CPI |
| CPI × SPI (composite) | EAC = AC + (BAC − EV) / (CPI × SPI) | Both cost and schedule pressure will affect remaining cost |
| Bottom-up | EAC = AC + new bottom-up ETC | Remaining work is re-estimated in detail |
Also useful:
ETC = EAC − AC
VAC = BAC − EAC
TCPI = (BAC − EV) / (BAC − AC) efficiency needed to finish on BAC
TCPI = (BAC − EV) / (EAC − AC) efficiency needed to finish on an approved EACWorked example (continuing Karachi Gateway Logistics)
Illustrative. BAC $2,000,000; at month 4: PV $800k, EV $700k, AC $850k, so CPI = 0.8235 and SPI = 0.875. Remaining work (BAC − EV) = $1,300,000.
| Method | Calculation | EAC |
|---|---|---|
| Budget rate | 850k + 1,300k | $2,150,000 |
| CPI | 2,000k / 0.8235 | ≈ $2,428,600 |
| CPI × SPI | 850k + 1,300k / (0.8235 × 0.875 = 0.7206) | ≈ $2,654,100 |
The range is roughly $2.15M to $2.65M. Which is most credible? That depends on causes:
- If the overrun came from a one-time event (e.g., a single re-excavation after an unexpected utility strike), the budget-rate method may be reasonable.
- If the cause is systemic (low productivity, under-estimated rates), the CPI method is more realistic. Experience across many projects suggests cumulative CPI tends to be fairly stable once a project is meaningfully underway, so claims of dramatic recovery deserve scepticism.
- If the team is behind schedule and will need to accelerate (overtime, extra crews), the composite method captures that pressure.
TCPI: the reality check
TCPI to BAC = 1,300k / (2,000k − 850k) = 1,300k / 1,150k ≈ 1.13. The remaining work must be done at 1.13 efficiency to hit budget, while the project has so far achieved 0.82. A jump from 0.82 to 1.13 is not plausible without a fundamental change. TCPI is one of the most effective tools for challenging optimistic forecasts in a steering meeting.
Bottom-up ETC: step by step
- List remaining scope by work package.
- Re-estimate each using current productivity, current prices and known changes.
- Add allowances for approved-but-not-yet-baselined changes and pending changes likely to be approved.
- Add remaining risk exposure (from the risk register or quantitative analysis).
- Compare the bottom-up EAC with the formula range. If it is below the most optimistic formula, you need a clear, evidenced explanation.
Forecast template
Control account: ______ Period: ______
BAC: ____ AC to date: ____ EV to date: ____ CPI: ____
Formula EACs: budget-rate ____ | CPI ____ | CPI×SPI ____
Bottom-up ETC: labour ____ materials ____ subcontract ____ other ____
Pending changes: ____ Risk allowance: ____
Selected EAC: ____ Rationale (2–3 sentences): __________
TCPI to BAC: ____ Credible? Y/N
Owner sign-off: ____Common mistakes
- Forecast equals budget until the final month ("hockey-stick" forecasting).
- Using the budget-rate method for systemic problems.
- Ignoring pending changes and claims.
- Presenting a single number with false precision instead of a range with rationale.
- Forecasting only at project level.
AI and forecasting
Machine-learning models trained on your organisation's historical projects can produce an independent EAC estimate and flag when the team's forecast is statistically unusual. Treat them as a third opinion alongside formulas and bottom-up estimates, and document which forecast was adopted and why.
Presenting the forecast
When you present an EAC to leadership, show the formula range, the selected figure, and the one or two assumptions that would move it most. For example: "Selected EAC $2.45M; range $2.15–2.65M; the main swing factor is whether steel productivity recovers after the new crew starts." This framing invites useful discussion rather than a debate about a single number.
Hands-on: forecast template in Excel
Named inputs: BAC, PV, EV, AC
CPI =EV/AC
SPI =EV/PV
EAC_budget =AC+(BAC-EV)
EAC_cpi =BAC/CPI
EAC_composite =AC+(BAC-EV)/(CPI*SPI)
TCPI_BAC =(BAC-EV)/(BAC-AC) meaningless once AC ≥ BAC: use TCPI_EAC
TCPI_EAC =(BAC-EV)/(EAC_selected-AC)
Credibility =IF(TCPI_BAC>CPI*1.1,"Budget not credible without a specific change","Plausible")
VAC =BAC-EAC_selectedKarachi check: CPI 0.8235, SPI 0.875; EACs ≈ 2.150M, 2.429M, 2.654M; TCPI to BAC ≈ 1.13.
Hands-on: the same checks in Python
def eac_range(bac, pv, ev, ac):
cpi, spi = ev / ac, ev / pv
rem = bac - ev
out = {
"CPI": cpi, "SPI": spi,
"EAC_budget_rate": ac + rem,
"EAC_cpi": bac / cpi,
"EAC_composite": ac + rem / (cpi * spi),
}
out["TCPI_to_BAC"] = rem / (bac - ac) if bac > ac else float("inf")
return out
for k, v in eac_range(2_000_000, 800_000, 700_000, 850_000).items():
print(f"{k:16s} {v:,.3f}" if v < 10 else f"{k:16s} {v:,.0f}")How to measure success
- Forecast accuracy: record each month's selected EAC and compare with the final cost; track the error by method.
- No period where TCPI to BAC exceeds achieved CPI by more than about 10% without a documented recovery action.
- Every EAC presented as a range, a selection and a rationale.
Key takeaways
- EAC = AC + ETC; use several formula methods as independent checks.
- Choose the method based on the cause of variance: one-off, systemic, or schedule-driven.
- TCPI shows the efficiency required to hit a target; a big gap from current CPI signals an unrealistic forecast.
- Final forecasts are documented judgments, ideally ranges, with rationale and sign-off.
Check your understanding
Quick questions to lock in the lesson. They don’t count towards your certificate.
Put it into practice
For the Karachi example, write a 3-sentence forecast rationale recommending one EAC and explaining why the other methods were not chosen.
Enrol for free to save your progress
Reading is always free. Enrol to keep your place, take the final assessment and earn a verifiable certificate.