Discovery and delivery

The best product work happens when engineering, product, and design solve the same problem together — not when a ticket arrives fully formed and your only job is to type. Marty Cagan's blunt test: if the first time engineers see an idea is sprint planning, you are on a feature factory, not a product team.

You will meet two failure modes. In the first, eng is a ticket factory: low agency, low learning, high resentment. In the second, eng ignores users and invents architecture for sport: high agency, low impact. The middle path — shared problem ownership — is where SDE2-shaped careers grow.

Roles, with overlap

Peers, not a hierarchy. A PM who dictates pixel-level solutions and an engineer who refuses to discuss users are both failing the same collaboration. Good teams argue in discovery and commit in delivery.

Practical boundary: if a debate is about user value or priority, lead with curiosity and customer evidence — the PM usually holds more of that context. If it is about feasibility, operability, or irreversible technical risk, lead with clarity and numbers. Many fights are really mixed. Split the question into those two layers before you dig in.

Cadence that keeps the trio aligned

If your team has none of these, propose one. You do not need permission to schedule a 25-minute "does this still match the goal?" with PM and design.

Mid-build check, worked: design added a "suggested folders" carousel that needs a new ranking service. Eng flags +5 days and a new on-call surface. The trio decides: ship without suggestions, instrument "create folder" clicks, revisit suggestions if creation rate is high. Argument took fifteen minutes because the metric and cost were on the table.

Common mistakes

• First sight of the idea at sprint planning. That is a feature factory.
• "That's a product question" to avoid thinking.
• Designing in a vacuum and throwing specs over the wall the other way (eng → PM).
• Skipping the post-launch metric check because "eng is done."

Check yourself

1. You first hear about a feature at sprint planning. What is missing? Discovery. Engineers should have been in the problem and options, not only the ticket.

2. A debate mixes "users want this" and "this needs a new service." What do you do first? Split value/priority from feasibility/risk. Different evidence, different owners.

Key takeaways

• PM owns why/what, design owns experience, eng owns how — with heavy overlap in discovery.
• Engineers belong in discovery and delivery.
• Argue in discovery, commit in delivery.
• Light rituals (huddle, spec, mid-build, launch, results) keep the trio honest.

Further learning

SVPG articles (Marty Cagan)A partnership dedicated to teaching best practices to product teams and product leadersSilicon Valley Product Group

Next, cut scope on purpose in cheaper product cuts.