To-Complete Performance Index (TCPI) Explained

What is the to-complete performance index?

The to-complete performance index is an earned value metric that states the cost efficiency required on all remaining work to hit a specified cost target. It is a ratio, not a percentage: a TCPI of 1.15 means the remaining work must deliver $1.15 of budgeted value for every dollar spent.

PMI defines it as the calculated cost performance index that must be achieved on the remaining work to meet a management goal such as the budget at completion or the estimate at completion. The US Department of Energy’s earned value guidance puts the same idea more plainly: work remaining divided by funds remaining.

Two things follow from that definition. First, TCPI is forward-looking — it says nothing about what has already happened, only what must happen next. Second, it is a test of a target, not a description of reality. You calculate it against a number someone has committed to, and the result tells you how hard the remaining work has to work to honour that commitment.

What are the two TCPI formulas?

There are two, and choosing the right one is the whole skill. Both use the same numerator — the budgeted value of work still to be done — and differ only in the denominator, which is the money still available.

• BAC is the approved budget at completion, after any formally approved changes.
• EV is the budgeted value of work actually completed.
• AC is the actual cost incurred for that same work.
• EAC is the current forecast of the final total cost.

The BAC version is the one that appears in exams and in most status reports. The EAC version is the one that matters in practice once the original budget is clearly lost, because it tests whether the recovery plan is credible rather than whether the original promise was wise.

A useful sanity rule: if TCPI(BAC) is 1.4 and TCPI(EAC) is 0.95, the team has accepted a real overrun, the forecast is honest, and the remaining work only needs to run at roughly plan efficiency. If both numbers are 1.4, either the overrun has not been acknowledged in the forecast, or an extraordinary recovery is being assumed. Neither survives contact with a steering committee for long.

How do you calculate TCPI step by step?

Five steps, no algebra beyond subtraction and division.

1. Fix the baseline. Use the approved BAC, including approved changes. Note the version and date.
2. Take earned value at the cut-off. Use the same progress rules you use for CPI, preferably 0/100 or 50/50 rather than optimistic percentages.
3. Take actual cost for the same scope. Costs must match the work measured, not the invoice cycle.
4. Choose the target. Original budget if it is still the commitment; EAC if a revised forecast has been approved.
5. Divide and write one sentence of context. The number without a driver is an argument waiting to happen.

Worked example: the six-month migration

A company is migrating 40 applications over six months with a BAC of $480,000. At the end of month three: EV is $210,000 and AC is $250,000, so CPI is 0.84.

• Remaining work: $480,000 − $210,000 = $270,000
• Remaining budget: $480,000 − $250,000 = $230,000
• TCPI(BAC) = 270,000 ÷ 230,000 = 1.17

The team has been delivering $0.84 of value per dollar spent. To finish inside the original budget, it must now deliver $1.17 — a 39% improvement in efficiency, sustained across the entire remainder of the project.

Now test the forecast. If the approved EAC is $617,347, then remaining funds are $617,347 − $250,000 = $367,347, and TCPI(EAC) = 270,000 ÷ 367,347 = 0.73. Against the revised forecast, the team can underperform plan and still land on the number. That is what an honest re-baseline looks like.

Worked example: late-stage distortion

A fit-out has a BAC of $1,000,000. At 90% complete, EV is $900,000 and AC is $980,000.

• TCPI(BAC) = (1,000,000 − 900,000) ÷ (1,000,000 − 980,000) = 100,000 ÷ 20,000 = 5.00

A TCPI of 5.00 does not mean the team must become five times more efficient. It means the denominator has almost disappeared: only $20,000 of budget remains to deliver $100,000 of budgeted work. The correct conclusion is that the original budget is gone by roughly $80,000 and the remaining scope is small enough that the final overrun is capped. Report the overrun, not the ratio.

Worked example: the under-budget case

A software release has a BAC of $300,000. At the halfway point, EV is $180,000 and AC is $150,000.

• TCPI(BAC) = (300,000 − 180,000) ÷ (300,000 − 150,000) = 120,000 ÷ 150,000 = 0.80

The team can spend $1.25 for every dollar of remaining budgeted value and still finish on budget. That headroom is worth protecting rather than spending: it is the reserve that absorbs the next surprise.

How do you interpret a TCPI result?

A TCPI of 1.00 means the remaining work must run at exactly the planned rate. Below 1.00 means you have slack; above 1.00 means you need to beat the plan. The useful interpretation depends on the gap between TCPI and the CPI already achieved.

Two habits keep the reading honest. First, compare TCPI to cumulative CPI, not to 1.00. A TCPI of 1.05 on a project running at 1.03 is a minor correction; a TCPI of 1.05 on a project running at 0.80 is a 31% step change in performance. The US Department of Energy treats a gap of 10 percentage points or more between CPI and TCPI as an early warning that the targeted budget may be unrealistic.

Second, look at both formulas together. The distance between TCPI(BAC) and TCPI(EAC) measures how much of the overrun the organisation has actually accepted. A wide gap is honesty. No gap, on a troubled project, is denial.

> Common mistake: quoting TCPI(BAC) to a team as a performance target. “We need 1.17 from here” is not a plan. It is a description of what the original budget now requires, and it usually needs a scope or funding decision rather than an exhortation.

Why does TCPI climb as a project progresses?

Because the denominator shrinks faster than the numerator. Remaining work falls in proportion to progress, while remaining budget falls in proportion to spending — and on an overrun, spending has been outrunning progress all along.

Take a project with BAC of $1,000,000 and a stable CPI of 0.85. At 20% complete, EV is $200,000, AC is $235,294, so TCPI(BAC) = 800,000 ÷ 764,706 = 1.046. At 80% complete, EV is $800,000 and AC is $941,176, so TCPI = 200,000 ÷ 58,824 = 3.40.

Nothing changed in performance. The index moved because the remaining budget became small relative to the remaining work. This is why late-stage TCPI values are dramatic and why a project can appear to “recover” simply by being re-baselined near the end.

The practical consequence: use TCPI as a decision tool between 15% and 70% completion. Before 15%, the numbers are dominated by mobilisation costs and are noisy. After 70%, the remaining work is small enough that the ratio exaggerates. In that late window, report variance at completion instead.

TCPI vs CPI: what is the difference?

CPI measures the efficiency you have achieved; TCPI measures the efficiency you now need. CPI is diagnostic and backward-looking. TCPI is prescriptive and forward-looking. The gap between them is the size of the turnaround being assumed.

A project with CPI of 0.92 and TCPI of 0.95 is managing a small overrun and has room to recover. A project with CPI of 0.92 and TCPI of 1.28 is asking for a 39% improvement in the remainder — a claim that needs evidence, not optimism.

When is TCPI meaningless?

Three situations break the metric, and all three are common.

When actual cost exceeds the budget. If AC is greater than BAC, the denominator in TCPI(BAC) is negative and the result has no interpretation. The budget target is already impossible; report the overrun and move to a revised EAC.

When scope has changed without a baseline update. TCPI compares remaining work against remaining money. If new scope was added without re-baselining, the numerator includes work the budget never covered, and the index inflates without telling you anything useful.

When earned value rules are unreliable. If progress is self-reported as a percentage on long tasks, EV drifts upward and TCPI drifts down. The metric is only as good as the progress rule behind it.

How do you use TCPI in a steering meeting?

Use it to frame a decision, not to assign blame. Three sentences do the work: here is the efficiency we have achieved, here is the efficiency the target now requires, and here is what would have to change for the gap to close.

A workable escalation pattern:

1. Report both TCPI values against BAC and against the current EAC.
2. State the required improvement as a percentage, not just an index: “1.17 against a CPI of 0.84 means a 39% improvement for the rest of the project”.
3. Name the mechanisms that could deliver part of it: descoping a work package, renegotiating a contract, resequencing to avoid overtime, replacing a underperforming supplier.
4. Ask for a decision on the residual: approve a revised EAC, release contingency, or reduce scope.
5. Record the decision and its date, so the next report tests the new target rather than reopening the old argument.

The department that handles this best treats TCPI as a reasonableness check on forecasts rather than a score. Contracting officers use it to test whether a supplier’s EAC is credible; project managers use it to test their own optimism. Both are asking the same question: is this number possible?

Does TCPI work in agile and hybrid projects?

It works if you can define a budget baseline and measure accepted work, and it is weaker where scope is deliberately emergent. Agile teams rarely have a fixed BAC and often forecast cost per sprint rather than per deliverable.

Where TCPI does help in agile is at the release or programme level. Fix the budget envelope for a release, measure earned value as the budgeted cost of accepted increments, and compute TCPI against that envelope. The result tells you whether the remaining increments can be delivered inside the money left — a question product owners and finance teams ask anyway, usually with worse arithmetic.

What TCPI cannot do is tell you whether the remaining scope is worth delivering. A project can hit its TCPI and still be a commercial failure. Pair it with the value case, not just the cost case.

Where teams get the inputs wrong

TCPI is a two-line calculation that fails on data hygiene, not on maths. The recurring problems are budgets that live in finance, progress that lives in a task board, and costs that arrive a month late. By the time all three are reconciled, the report describes a project that no longer exists.

Keeping planned value, earned value and actual cost against the same work packages in one workspace is the difference between a monthly calculation and a monthly argument. Doitify holds tasks, budgets, milestones, dependencies and cost reports in one place, so TCPI and CPI come from the same dataset on the same cut-off date. We build Doitify, so treat that as a product claim rather than independent advice — the discipline of one cut-off date matters more than the tool.

Conclusion

TCPI is the honesty check on a cost forecast. It takes a target someone has committed to and converts it into the efficiency the remaining work must deliver — which is either a plausible number or a clear signal that the target has to change.

Use it with three habits: report both the BAC and EAC versions, compare the result with the CPI actually achieved rather than with 1.00, and translate the index into a percentage improvement before anyone is asked to deliver it. Then act on the gap between 15% and 70% completion, while there is still scope, schedule and contingency to trade.

The projects that recover are not the ones with the most optimistic forecasts. They are the ones where someone calculated what recovery would require, said it out loud, and changed the plan before the arithmetic became impossible.

If this post on to-complete performance index was helpful, you might also enjoy Why Projects Miss Deadlines and How to Fix It and Project Manager vs Program Manager: What’s the Difference?.