Why Your Project Backlog Never Ships — And What Actually Fixes It

Every company has a backlog. A list of projects that everyone agrees are worth doing, that have business value attached, that someone has already thought through — and that simply never get started. The usual explanation is that there isn't enough budget, or enough engineers, or enough clarity. But if you look closely at where projects actually get stuck, a different and more uncomfortable answer emerges.

The backlog doesn't grow because you're out of ideas. It grows because you're out of coordination capacity.

Coordination is the real constraint

Think about what it actually takes to move a single project from "approved" to "shipped." Someone has to break the work into tasks, sequence them, assign owners, and put them on a board. Then — and this is the part nobody accounts for — someone has to keep that picture accurate as reality drifts. Cards get stale within days. A dependency slips and three downstream dates are now wrong. A contributor goes quiet and nobody notices until the weekly sync.

That ongoing maintenance is not glamorous, but it is relentless, and it is what consumes the majority of a project manager's day. Industry analyses of knowledge-worker time consistently find that a large fraction of project-management effort goes not to planning or strategy, but to administrative upkeep: updating statuses, grooming boards, and chasing people for information.

Here's the trap. Your engineering capacity can scale. Your design capacity can scale. But your coordination capacity is capped at the number of project managers you can afford to hire — and each of those people can only hold so many projects in their head at once before quality degrades.

Why hiring more PMs doesn't solve it

The instinctive fix is to hire. But adding project managers has three problems. It's expensive: experienced PMs command significant salaries, and the coordination overhead they relieve is, by definition, lower-leverage work. It's slow: a new PM takes months to onboard into your context, tools, and relationships. And it scales linearly: two PMs can run roughly twice as many projects as one, while the projects you want to run grow far faster than that.

So the backlog persists. Not because the projects aren't valuable, but because the cost of coordinating them exceeds the coordination capacity you have. Projects sit in the queue, lose their urgency, and quietly die.

Your engineering capacity scales. Your coordination capacity is capped at the number of project managers you can afford.

The work is more automatable than it looks

The good news is that coordination work has a particular shape: it is structured, rule-based, and repetitive. A Kanban board is, quite literally, a state machine. Tasks move between defined columns according to defined rules. Follow-ups happen on defined cadences. Escalations fire on defined conditions. This is exactly the kind of work that modern AI handles well — not the creative judgment of deciding what to build, but the disciplined execution of keeping everyone aligned on getting it done.

McKinsey's research on generative AI in operations suggests a meaningful share of routine coordination and knowledge work can be automated. Gartner has projected that by 2030 the majority of project-management tasks will be handled by AI. The direction of travel is clear: the administrative layer of project management is becoming a solved problem.

What "breaking the constraint" looks like

If coordination is the constraint, then relieving it — rather than adding more humans to absorb it — is what unlocks the backlog. Concretely, that means a system that can:

• Maintain an accurate board without anyone manually updating it.
• Follow up on slipping tasks the moment they slip, not at the next weekly meeting.
• Escalate genuine risks with context, so leaders intervene before deadlines are missed.
• Do all of the above across many projects at once, around the clock.

When the coordination ceiling lifts, the math changes. The same team of project managers can oversee far more projects, because they're spending their time on judgment and exceptions instead of status maintenance. Projects that would have languished in the backlog get a coordinator on day one.

From a system of record to a system of execution

Traditional tools — the boards and trackers most teams use — are systems of record. They store the state of your projects, but they rely entirely on humans to keep that state current. They are passive. They wait for you.

The shift that actually clears the backlog is moving to a system that does the coordinating itself: that reads what's happening, updates the record, and pushes work forward without waiting to be told. That's the premise behind Vertically — an AI project manager that runs your projects over email, so the constraint that's been quietly killing your backlog finally lifts.

More projects. Less hiring. It starts with recognizing that the bottleneck was never your ambition. It was your bandwidth.