New engineering executives often arrive with a playbook from their last company and a mandate to fix things. The first 90 days reward the opposite instinct: listen deeply, diagnose carefully, and make a small number of high-leverage changes.

The goal is not to prove you are smart. It is to build enough trust and context to make good decisions for the next two years.

Days 1–30: build the map

Meet everyone who touches engineering: direct reports, product partners, sales, support, finance, and a meaningful sample of engineers. Ask the same five questions:

  • What is working that we should protect?
  • Where do you lose the most time?
  • What decision are we avoiding?
  • Who is doing invisible work that holds this place together?
  • What would you do in my first month?

Read the last two quarters of incident reviews, planning documents, and exit interviews. Resist the urge to reorganize. You do not yet know which problems are structural and which are seasonal.

Days 31–60: diagnose the system

Move from individual conversations to system-level patterns. Map the flow of work from idea to production. Where does it queue? Where does it get reworked? Which teams are consistently waiting on which others?

Assess four dimensions: delivery, reliability, talent, and operating rhythm. For each, write down what you observe, what you infer, and what you still need to learn. Separate facts from hypotheses. Share the draft diagnosis with your leadership team and invite correction.

Trust is not built by having the answers. It is built by being honest about what you do not yet know — and then learning in public.

Days 61–90: choose a narrow agenda

Pick three priorities, not ten. A credible first agenda usually includes one delivery problem, one organizational problem, and one reliability or platform constraint. Each priority needs a measurable outcome, a named owner, and a review date.

Communicate the agenda in plain language: why these three, what will change, what will not change yet, and how people can contribute. Explicitly name what you are choosing not to tackle this quarter. This protects focus and reduces the anxiety of an endless priority list.

Build your operating rhythm

Establish the minimum cadence you need: weekly direct reports, biweekly skip-levels, monthly engineering all-hands, quarterly planning. Keep the first versions light. You can add structure later; removing it is much harder.

What to avoid

Do not reorganize in week one. Do not cancel a project to signal decisiveness. Do not promise a timeline you have not validated. And do not confuse being liked with being trusted. The first is pleasant; the second is the foundation of everything that follows.

By day 90, you should have a shared diagnosis, a narrow agenda, and a team that believes you understand the place. That is a strong start.