You don't have a PMO. You still have a portfolio.

Thursday evening, you message six people.

How's the migration going. Did the enterprise pilot get its data. Are we still shipping the redesign this month. What happened with the vendor. Do we need to move the hire. Is the compliance thing done.

Four reply that night, two reply Friday, one of those replies with a question rather than an answer. You read all of it, hold it in your head, and by Monday you have a picture of the business that exists in exactly one place: you.

Nobody calls this portfolio management. That's a term from companies with a governance function and a template library. But that's what it is. Fifteen things are in flight, they compete for the same people and the same money, some of them block each other, and one person is tracking all of it by memory and Thursday-night messaging.

You are the integration layer

Here's the thing to notice. Every one of those fifteen efforts is documented somewhere. Engineering has tickets. Sales has a pipeline. The redesign lives in Figma and a doc. Finance has a sheet. The vendor conversation is in your inbox.

What doesn't exist anywhere is the join. Nobody has ever written down that the enterprise pilot depends on the migration, which depends on the hire you deferred in March. You know it. You know it because you were in all three conversations. That knowledge lives in your head and it is not written down in any system, which means the company can't reason about it without you in the room.

That works, for a while. It works when there are five efforts and you were personally in every meeting. It stops working somewhere around a dozen, and it doesn't stop gradually — it stops the first time two things collide that you'd have caught six months earlier, and you find out three weeks after the collision.

Why hiring a PM doesn't fix it

The reflex at this point is to hire someone to own the tracking.

It helps. It also doesn't solve what you think it solves, because the problem was never that nobody was writing things down. The problem is that the information is scattered across systems that don't talk, and a new hire arrives into exactly the same scattering. What you've done is add a person whose job is to manually assemble a picture and tell it back to you — which is genuinely useful, and is also a report with a two-to-nine-day lag on it, produced by someone who wasn't in the room for the last three years of context.

You've bought yourself a reporter. You still don't have visibility. The distance between the two is the evidence gap: what your systems already know, minus what anyone can actually see.

Three questions your tools should answer and don't

Strip away the ceremony and an operator needs three things answered, continuously.

What changed this week that I'd have wanted to know about? Not everything that changed. The things a reasonable person would flag — a date that moved, a dependency that broke, a decision made without you that has consequences beyond its team.

What's about to slip, and what does it block? Slippage in isolation is a scheduling issue. Slippage that lands on the critical path of two other efforts is a business issue. The second is the one you need, and it's the one nobody can see, because the dependency was never written down anywhere except your head.

Where are money and people actually going, versus what I said mattered? Every company has a gap between stated priorities and revealed ones. It shows up in where the senior engineers' weeks go, not in the strategy deck.

Now try answering those three from your tools this Thursday. Not from your team — from the systems. Time yourself.

You won't get through the first one, because Jira knows about tickets and the calendar knows about meetings and neither of them knows about the other, and the thing you needed was in the join.

Try this Monday

Write down the fifteen things currently in flight. Next to each, write the one thing that would have to go wrong for it to matter to you personally.

Then, for each of those fifteen risks: could you tell today whether it's happening?

The ones where the answer is I'd have to ask someone are your exposure. Not because your people are unreliable — because the only detection mechanism you have is a person choosing to raise it, and people raise things late for entirely human reasons.

Next in this series: the risk was visible three weeks before anyone escalated it.