Who does project management work when there are only product managers?

Why and where is project management growing while tech cuts layers

Despite our tech-centric perspective, project management is not shrinking. The Bureau of Labor Statistics counts 1,094,300 project management specialists in the US and projects 7% growth through 2035, which it describes as much faster than the average for all occupations. Construction employs 22% of them against 5% for finance and insurance, while professional, scientific and technical services take the largest share at 28%, which is where most of the consulting and engineering services work sits. Indeed's posting index tells the same story from a different angle, with construction postings running 23% above their pre-pandemic baseline and industrial engineering 15% above, while software development sits 25% below.

Under the US infrastructure act, about 74% of available funding has been legally obligated and only about 45% has actually been spent, and the gap between committed and delivered is measured in hundreds of billions of dollars of work that somebody has to sequence, permit, staff and land. PMI, which has an obvious interest in the answer and should be quoted with that in mind, projects demand for up to 29.8 million additional project professionals globally by 2035, with construction, manufacturing, IT services and healthcare under the most pressure.

Meanwhile Amazon cut 14,000 corporate roles in October and another 16,000 in January, with the company explaining both rounds as reducing layers and removing bureaucracy, and Google's VP of people analytics told an internal meeting the company had 35% fewer managers running small teams than it did a year before. It’s the same job but in two different industries headed in opposite directions, and I feel like the usual explanation for the split, which is that tech got smarter while government stayed bloated, is mostly flattering noise.

What deterministic work gives project managers that discovery work does not

Think about what the growing industries have in common. A substation, a satellite, a runway extension, a benefits system with a statutory go-live date. The outcome is specified before anyone starts, the requirements are fixed from outside by physics or regulation or contract, the sequence genuinely matters because you cannot pour the second floor first, and the difficulty sits in coordinating dependencies and risk across many parties over several years. That is the home ground of the discipline which is why the aerospace and defense industry employing over 2.23 million people in the US keeps hiring for it.

Now think about what a software product organization actually works towards. The outcome is a hypothesis. The requirements are the thing you are trying to discover, and finding out you were wrong in the 6th week of work counts as a success rather than a variance. This is the argument for running discovery and delivery in the same cadence instead of sequencing one behind the other. Ultimately a role built on tracking a plan has a hard time when the plan is the least stable artifact in the room, which is why we keep pushing teams toward roadmaps organized around outcomes rather than dates. Technology organizations stopped having projects and moved towards product. The job of project management followed the projects out the door.

Who does the project management work in tech now that scrum masters are out of fashion

We did try to solve this, and we solved it by inventing a container. In January 2023 Capital One eliminated 1,100 roles in one move, the entire agile job family. In their defense, they said, “the agile role had been critical to earlier phases of the transformation, and as the organization matured the natural next step was to integrate agile delivery processes directly into core engineering practices.” Which is a polite way of saying the work went to the engineers and the product managers. The pattern will look familiar to anyone who has watched a company bolt a new practice onto a framework instead of redesigning the work underneath it.

Product Focus surveyed 677 product people across 40 countries this past winter and found that 60% frequently spend time on unplanned activities, only 25% spend most of their time on strategic work, and 41% feel their role lacks clear definition. Teams whose responsibilities are clearly defined hit their deadlines 78% of the time against 48% for teams where they are not. The survey does not attribute any of that to absorbed project management work, and I want to be careful that the connection is my inference rather than their finding, but it lines up with what I see in the rooms I work in. Meanwhile the managers who might have picked up the slack have less room than they used to, with Gallup's span of control data across roughly 92,000 teams showing average direct reports rising from 8.2 in 2013 to 12.1 last year.

The work did not evaporate when the title did. Somebody still reconciles the dependency between two teams that share a service, somebody still writes the update the executive sponsor reads on Thursday morning, somebody still notices that the vendor contract expires three weeks before the migration finishes, and in most of the organizations I work with that somebody is a product manager who was hired to do a different job or an engineering manager who now has twelve reports instead of six. I suspect the original mistake was handing the work to a role instead of building it into a practice, because the scrum master was a container we invented for coordination and when budgets tightened the container was the easiest thing in the room to cut.

The teams I see come through this well are the ones that made the coordination work explicit, owned inside the team, and tied to an outcome somebody is actually accountable for, which is most of what we end up teaching when a client asks us to fix delivery and we find the real problem is that nobody knows where the team boundaries sit or who owns the seams between them. In many ways that is just the old argument about measuring outcomes rather than story points showing up in a new costume.

Worth asking on your own team this week: who wrote the last status update your executive sponsor actually read, how long did it take them, and what did they not get to instead?

Previous
Previous

What will product management look like in 2027?

Next
Next

Training 200 to 2000 product managers in the AI era requires a human touch