Where does engineering time go?

Someone above me asked how much of engineering goes into new product and how much goes into keeping things running. I want a real answer, not a feeling. ## Context - Scope: [all squads | squad names] - Period: [last 90 days], compared with the 90 days before it - Unit: [issues | story points]. If both are available, run both. - The split leadership believes it is funding: [unknown | for example 70% roadmap, 30% everything else] - If I left anything in [brackets], keep what reads as a default (a period, a number of days, the first of several options) and ask me for the rest in one go, before pulling any data. - Write everything, the deliverable included, in the language of this prompt. ## Pull this data from DevStats (both periods, same filters) 1. Allocation — roadmap, unplanned, planned maintenance, unplanned maintenance 2. Engineering Investment — new things, improve things, keep the lights on, productivity 3. Investment Profile — features, bugs, enhancements, hotfixes, others 4. Work Breakdown — new code vs refactor vs rework 5. Throughput — volume, for context only 6. Issue List resolved in the period with issue_type bug and then hotfix, sorted by cycle time descending, limit 10 each, fields key, summary, project, cycle_time_formatted. These are the largest unplanned items. ## How to analyze - The three reports cut the same work in different ways. Allocation answers planned vs unplanned (predictability). Investment answers new vs maintain (strategy). Profile counts types. Use each for its own question; never add them together. - Counting by issues overweights small bugs. Counting by story points overweights large features. If both are available, show both and say which one is closer to effort for this team. - There is no universally right split. A young product at 70% new is normal, and so is a mature platform at 40%. What matters is the gap between the real split and the one leadership thinks it is paying for. - Unplanned work above ~30% means the roadmap is a wish. Find what feeds it: hotfixes and bugs are quality, "others" is work nobody labeled. - A large unlabeled share means part of the picture is blind. Say how much, and do not draw conclusions from that part. - Rework share rising means we are paying twice for the same feature. - Differences between squads can be by design. A platform squad at 80% keep-the-lights-on is doing its job. - Make it tangible without inventing costs: "of every ten engineering days, three go to work nobody planned". ## Deliver A one-page memo with one chart: 1. **The answer** — one sentence in the form "of every ten engineering days..." 2. **Chart** — stacked bars, current vs previous period, roadmap vs the rest 3. **What feeds the unplanned work** — the top five sources or items, with evidence 4. **The gap** — real split vs the split leadership expects, in points 5. **Two levers** — the two changes most likely to move ten points back to roadmap, and what each one costs 6. **Confidence** — how much of the work is unlabeled and what to fix in the tracker before the next read