All posts

Who Should Decide What Engineers Build Next?

A product team at a whiteboard deciding what to build next, with a stack of stakeholder tickets set aside.

The product team decides what engineers build next: product, design, and engineering together, working from outcomes and evidence. Executives set the outcome, the strategy, and the constraints, not the queue. Stakeholders bring constraints and customer access. If the next build is decided by whoever escalated last, nobody is actually deciding.

Whoever escalated last is not a decision process

If the next engineering week is decided by whoever escalated last, you do not have product management. You have a ticket router with a salary.

Who should decide what engineers build next? The people accountable for the outcome—working as a team—using evidence, constraints, and an explicit strategy. Not a committee of drive-by requests.

Empowered decisions

Cagan’s product manager contribution puts value and viability on the PM, usability on design, feasibility on tech. The next build should emerge from that triad solving for a problem, not from a stakeholder wish list passed through a proxy.

Two-in-a-box fails here too: if the person who understands customers never sits with engineers, the “decision” is already a handoff.

Cutler’s outcomes framing asks whether the way you work produces better outcomes. A stakeholder queue optimizes for political throughput. That is an outcome—just not usually the one on the slide.

Good vs bad decision rights

Good decision rights Bad decision rights
Leadership sets problems, constraints, and success measures Sales writes the roadmap in the CRM notes
The product team chooses solutions and sequencing inside that box Executives reorder the sprint mid-week without changing capacity
Stakeholders are experts on constraints, not a second backlog The "PO" forwards every request to stay safe
Engineers can challenge feasibility and propose better approaches Engineering estimates a list they were not allowed to shape
"No" and "not now" are normal, documented, and durable Every request is "yes, next sprint"

The same rule applies outside the company. If a vendor decides what your team builds next, the vendor roadmap is not yours, and neither is the queue.

Practical rule for the next item

Before something enters build, someone accountable can answer:

  1. What outcome does this serve?
  2. What evidence says this is the right bet now?
  3. What will we stop or delay to fund it?
  4. How will we know it worked?

If those answers are missing, the item is not ready—no matter how loud the request.

Connection to earlier posts

We argued backlog management is not product strategy, discovery must sit beside delivery, and product teams own outcomes—see what a product manager should own. Decision rights are how those ideas become Tuesday’s commit.

For companies buying software help, this is why consulting often precedes custom software. Encoding a broken queue into a nicer tracker still leaves you with a queue. Research before another screen is the same discipline at the idea layer.

If engineers are a shared service for every idea, talk to us.

Frequently asked questions

Should executives decide the roadmap?

Executives should decide strategy and outcomes. Solution selection belongs with the team closest to the evidence—inside those constraints.

Where do stakeholders fit?

Stakeholders fit as sources of constraints and customer access, not as an unordered second backlog.

What if legal or compliance must dictate work?

Legal or compliance requirements count as part of viability. Sequence them deliberately instead of letting them arrive as surprise interruptions.

Can engineers decide alone?

Engineers can decide alone on pure feasibility and tech risk. Value and viability still need product judgment, or engineering becomes an expensive hobby shop.