Why PMOs Fail Even When Every Project Is Green

Most program reviews look the same. A slide with a row for every project, a column for schedule, budget and scope, and a wall of green. Everyone nods. The meeting ends early.

And somewhere in that room, the program is quietly failing.

I don’t think that happens because anyone is lying. It happens because we measure projects and assume the program is the sum of them. It isn’t. A program is the projects plus everything that has to happen between them, and that second part is exactly what a project status report is built to leave out.

The Watermelon Problem

There’s an old name for this: the watermelon report. Green on the outside, red on the inside.

Status reports drift toward green for ordinary, human reasons:

Amber gets punished. The project manager who reports a problem early gets the extra meetings, the escalation and the scrutiny. The one who reports green gets left alone. People learn fast.
Green is rounded. “We’ll recover that slip next month” becomes green this month. Do that a few times and the color stops describing reality and starts describing hope.
Status measures the plan, not the outcome. A project can hit every milestone in its plan and still deliver something nobody can use. The report asks “are we doing what we said?” and never “is what we said still right?”
Each report covers one scope. Every project manager reports on their own project, honestly. Nobody is asked to report on the space between projects, so nobody does.

None of that requires bad intent. It’s just what happens when the reporting system rewards calm over truth. In my personal operating code I wrote that I won’t let optimism replace evidence. A wall of green is often optimism with a nice format.

Where It Hides: The Seams

Here is a simple, made-up example. Picture a program with four projects: a new business system, a data migration, a process redesign and staff training. Each has a capable lead and a solid plan.

The system project is on schedule, assuming the migrated data shows up in the format it expects.
The migration project is on schedule, delivering data in the format it agreed to six months ago.
The process team is on schedule, designing workflows around features that were descoped from the system last quarter.
The training team is on schedule, building courses for a go-live date the other three no longer quite believe in.

Four green rows. One red program.

Every one of those problems lives at a seam: a handoff, a shared assumption, a dependency, a decision that touches more than one project. Seams don’t belong to any single project, so they don’t appear on any single report. They belong to the program, and in most programs nobody owns them.

That’s the real job of a PMO or a program office. Not collecting status, but owning the seams.

The SEAMS Check

Project status tells you whether the parts are on track. To know whether the program is healthy, I look at five signals instead. Each one is something a project report won’t show you on its own.

S: Shared Dependencies

List every handoff between projects: who gives what to whom, and when. Then ask both sides to confirm it. A dependency is only healthy when the giver and the receiver describe it the same way.

The tell is a dependency that’s green on the giving side and unknown on the receiving side. That’s a surprise with a date on it.

E: Evidence of Integration

When did the pieces last work together? Not each component tested on its own, but an end-to-end run, a pilot, a walkthrough with real users. Integration is where assumptions meet each other, and the longer you go without it, the more expensive the meeting.

A program that has never been seen working as a whole has no evidence that it works. It has only a plan.

A: Aging Decisions and Risks

Track how long things wait. How many days has the oldest open decision been sitting? How long has the top risk been on the register without its mitigation moving?

Age matters more than count. A program with many fresh risks is being managed. A program with a few old ones is being avoided. Decisions that age are usually decisions that cross a seam, where nobody is sure whose call it is.

M: Mission Outcome

Go back to why the program was funded: the benefit, the capability, the change it was supposed to deliver. Then ask whether that’s still reachable, given everything that’s been descoped, deferred and “moved to phase two.”

Programs rarely fail all at once. They fail one reasonable trade-off at a time, until every project finishes and the outcome quietly didn’t make the cut.

S: Signal Honesty

Finally, look at the reporting itself. When did a project last turn amber before there was a crisis? When was bad news last delivered early and met with help instead of heat?

If the answer is “never,” the colors aren’t data. They’re a reflection of the culture. A healthy program has some amber in it most of the time, because real work is uncertain and people feel safe saying so.

Putting It to Work

You don’t need a new tool for any of this. You need a different agenda.

Give the seams an owner. Someone at the program level owns the dependency list, the integration plan and the decision log. If that sounds like the PMO’s job, that’s the point.
Review the seams, not just the rows. Spend the first part of every program review on the five signals. Project status can come second, or go in the pre-read.
Reward early amber. When someone raises a problem early, the response should be help, not a hearing. Do it visibly. People watch what happens to the first person who tells the truth.
Write it down. Seams owned in someone’s head disappear when that person is out. Make ownership visible, as I argued in Build for Absence.

This also changes what we ask of project managers. As I wrote in Digital Transformation: The Pivotal Role of Project Managers, the job is leading change, not just delivering scope. A project manager who flags a broken handoff with the project next door is doing more for the program than one who reports a perfect green.

It’s the same pattern I described in The Hidden Cost of Cloud Waste: the work succeeds by its own measure while the thing you actually cared about slips away.

What to Do Monday

If you sponsor a program, you don’t have to rebuild its reporting to see past the green. Ask three questions at your next review:

Which handoff between projects worries you most, and who owns each end of it? If nobody can name one, nobody is looking. If they name one but can’t name both owners, you’ve found your first seam.
What decision is this program waiting on, and how long has it been waiting? The answer tells you where authority is unclear, and how much schedule is quietly burning while it stays that way.
When did we last see the pieces work together, and what broke? “We haven’t yet” is an honest answer. It just isn’t a green one.

Listen less for the answers than for the pause before them. A team that answers fast is watching the seams. A team that has to go and find out has been watching the rows.

The Bottom Line

A program isn’t a list of projects. It’s a set of promises the projects make to each other, and those promises are where programs live or die.

Green projects don’t make a green program. Owned seams do.

Measure the seams, give them an owner, and make it safe to report amber. The colors will get less comfortable, and a lot more true.