- May 30
Half your engineers' time is on something other than active engineering
- Niklas Sundgren
- Field Notes
I have asked senior R&D leaders the same question many times. What are your engineers actually spending their time on?
The honest answer is usually some version of "I have a sense, but I cannot prove it." Project-coded time accounts for part of the picture. Some leaders run timesheets. Most do not. None of them can describe with confidence what the function actually does on a typical week.
The number worth knowing is 50%.
Bain's 2023 research on engineering and R&D talent gaps documents the picture across aerospace and defense: engineers spend roughly 50% of their time on active engineering work. The rest is consumed by rework (30%+) and lower-value tasks (40%+, with category overlap).
That is the industry-typical baseline. Most R&D leaders have not measured their own function against it. The published utilization numbers do not show the gap, because the gap is in the categories the published numbers do not track.
Invisible work is the term for what fills it. Undocumented support to other functions. Informal commitments to peers in sales or product. Internal services nobody chartered. Technical maintenance on systems nobody owns. Half-staffed projects that consume engineer attention without showing up in the portfolio. These accumulate. The function ends up doing more than its capacity suggests because the visible portfolio is only part of the load.
The R&D leader operates with a private estimate that diverges from the published picture. The published utilization says there is room for one more project. The private estimate says we are 30% over capacity already. The private estimate cannot be defended in a resource conversation without data the function is not collecting.
The instinct, when this gap surfaces, is to install timesheets. This is usually wrong. Heavy time-tracking systems reduce trust, encourage gaming, and produce data that engineers and managers stop believing within months. The function then has the worst of both: invasive measurement and unreliable numbers.
The alternative is light governance. Sampling-based measurement, not continuous tracking. Periodic structured conversations rather than weekly logs. R&D-specific categories that map to how the work actually happens, not generic timesheet codes. The goal is enough visibility to make resource decisions on data, not enough overhead to damage the trust that makes engineers honest about how they spend their time.
What changes when you build the picture:
Resource conversations get settled on data instead of opinion. The published utilization stops being the starting point of an argument and starts being a defensible claim.
Hiring decisions get anchored to a real baseline. The function knows what it is buying capacity for.
Invisible work gets named. Whether the organization wants to keep it (in which case it gets formally resourced) or stop it (in which case the engineers absorbing it get their time back) becomes a decision rather than a drift.
Portfolio commitments stop assuming engineers are 100% available for portfolio work. The math starts from the actual baseline.
If your function's published utilization numbers are not numbers you would defend in a CFO conversation, the function is operating on capacity claims that nobody fully believes. That includes the leader making them.
If this is the visibility gap your portfolio decisions have been operating without, let's talk.