CPI and SPI: How to Read Them in a Steering Committee
In a steering committee, CPI and SPI aren't "presented" — they're interrogated. The difference between a meeting that decides something and one that just listens to a report lies in how those two numbers are read — and what happens when they disagree.
This article assumes you already know where CPI and SPI come from. If not, start with What Is Earned Value (EVM) and come back. Here we go one step further: interpreting them in front of a committee.
The two numbers, one sentence each
- CPI (EV / AC) — how much value you deliver for every dollar you spend.
0.90means that for every dollar spent, you produce 90 cents of planned work. - SPI (EV / PV) — how much work you've completed against what the plan said you'd have.
0.90means you're 10% behind schedule, measured in value.
They share the same conceptual denominator — EV, what's actually been produced — which is why they're read together. A CPI without its SPI next to it says nothing useful.
The reading rule: never one without the other
The correct reading is a two-axis matrix. Place the project in the quadrant that fits depending on whether each index is above or below 1:
| CPI | SPI | What's happening | What to ask |
|---|---|---|---|
| ≥ 1 | ≥ 1 | Performing and on time. The healthy case. | Is progress being measured properly, or is it over-reported? Is there slack that could be reallocated? |
| ≥ 1 | < 1 | The money performs, but delivery is late. Usually a staffing gap or blocked dependencies, not a cost problem. | Is the critical path identified? Is the delay recoverable, or does the finish date need to move? |
| < 1 | ≥ 1 | On time, but every unit of progress costs more. Rework, overtime, poorly estimated scope. | Are you buying schedule with overrun? Will the budget hold to the end at this rate? |
| < 1 | < 1 | Overrun and delay. The project needs a decision, not another action plan from the PM. | What's the EAC and the projected date? Cut scope, add budget, or stop? |
The quadrant matters more than the exact value. A project that's "performing but late" has a different problem — and a different owner — than one that's "on time but expensive."
The threshold that matters isn't 1.00
No real project sits exactly at a CPI of 1.00. Reacting to a 0.98 is noise; ignoring a 0.98 for three straight weeks is negligence. That's why the committee needs tolerance bands agreed in advance, not an improvised judgment call at every meeting:
| Band | CPI / SPI | Committee action |
|---|---|---|
| Green | ≥ 0.95 | Logged. Not debated. |
| Amber | 0.90 – 0.95 | The PM explains the cause and the plan. Reviewed at the next committee. |
| Red | < 0.90 | Decision point: cut, reinforce, or close is put on the table. |
The exact cutoffs are up to each organization — an R&D project tolerates more variance than a regulatory one. What isn't negotiable is setting them beforehand, so the status color is arithmetic, not politics.
Example. An 18-month, $8M program. Month 9: CPI = 1.03, SPI = 0.84. Reading: the money performs, but the program is 16% behind. The PM proposes "catching up over the next two sprints." The committee instead asks for the SPI over the last three months: 0.91 · 0.87 · 0.84. It isn't catching up — it's slowly sinking. The decision isn't to wait another sprint; it's to move the committed date and notify the client now, with four months of runway, instead of in month 17.
The SPI trap near project closeout
SPI has a known flaw: as closeout approaches, it converges toward 1.00 even if the project is running late. The reason is mathematical — once nearly all the work is done, EV approaches total PV and the ratio climbs, even if that last task is two months behind.
In practice: in the final third of the project, stop watching SPI and watch two more honest things instead:
- SV in time — schedule variance expressed in days or weeks, not dollars — calculated against the baseline.
- The projected finish date against the committed one. If the final milestone keeps slipping, an SPI of 0.98 won't save you.
A single data point isn't a trend
The most expensive governance mistake isn't misreading the number — it's reacting to a single point. A given week's CPI can drop because a large invoice landed, or rise because a payment was delayed. What governs a committee is the direction:
- A CPI of 0.88 that's been climbing for three months is a project correcting itself. Let it work.
- A CPI of 0.96 that's been falling for three months is a project that will be red next month. Act now.
That's why a committee report shouldn't show today's CPI and SPI, but their curve over the last 8–12 weeks. The single value is the headline; the slope is the news.
The three-question script
When a project enters amber or red, the committee doesn't need a 12-point action plan. It needs answers to three questions:
- What's the root cause, in one sentence? If the PM can't summarize it, they don't know it yet.
- What does the EAC / date project if nothing changes? Today's expected cost and finish, not an aspiration.
- What decision do you need from us? Budget, scope, priority, or people. If the answer is "none," the project shouldn't be taking up committee time.
PMOvio gives you CPI, SPI and their trend
The curve over recent weeks, the EAC, the projected finish date, and the PMBOK status color — per project and across the whole portfolio, no spreadsheets. 15 days free, no card required.
See my portfolio →