Product Trio Framework

Unclear priorities in planning turned into design rework

How product, engineering, and UX started planning together at Delta.

Role: UX Advisory Architect. I introduced the format and tracked the results.
Timeline: The first squad started in October 2024. I measured results across 2025 Increments 1–4.
Team: Product Manager, Technical Architect, and UX designer. Technical Product Engineers (Delta’s business analyst role) ran the sessions.
Tools: Microsoft OneNote, Microsoft Loop, Agility

55→100%The first squad’s delivery predictability over three increments. The target was 80%.

3+1Three squads used the format, and nobody required it. A fourth started its own version.

SharedThe roadmap moved out of individual PM slide decks and into joint planning.

Delivery predictability by increment

Delivery predictability is one of Delta IT’s two key KPIs. It measures how much of the work a team commits to gets delivered. I took the first measurement in 2025 Increment 1.

The first squad passed the 80% target in the third increment and stayed above it in the fourth. 2025 Increments 1–4.

Each squad after the first heard about the format from a colleague. Nobody required it.

Problem & stakes

The squads did not agree on priorities, and late changes came back to design as rework.

The catering team was new, and its squads did not agree on priorities. It was not clear who owned the roadmap, and delivery was hard to predict. The team worked across Atlanta, Minneapolis, and other locations, so each planning cycle started with a two-day session just to get everyone caught up.

When I first measured predictability, in 2025 Increment 1, the first squad was at 55%. It delivered just over half of the work it had committed to.

As UX Advisory Architect, I had no authority over Product or Engineering. But when priorities changed late, the work came back to design as rework. So I decided to fix the problem in planning, before it reached design.

Process

I set it up to keep running without me.

I introduced the Product Trio, a format popularized by Teresa Torres. The Product Manager, Technical Architect, and UX designer plan together on a regular schedule.

Technical Product Engineers ran the sessions

I chose not to run the sessions myself. I was the only UX resource across several squads and could not attend all of them, so the format had to work without me.

From one session a week to two

The first squad started with one session a week in October 2024. As the list of tracked items grew, one session was not enough, so the squad moved to two.

Keeping the sessions on upcoming work

The main problem was that sessions kept drifting into current defects. The facilitators opened each session by saying the agenda was upcoming work only, and they steered the discussion back when it drifted.

Trade-offs

With someone else running the sessions, I had less control over them.

I did not control how each session was run. I accepted that, because a format that needed me in the room could only go as far as the squads I could attend.

I also tracked predictability only for the first squad. For the other squads, I can show that they adopted the format, but I have no predictability data.

Outcome

Other squads picked up the format without being asked.

Predictability: 55%, 66%, 100%, and 88% across 2025 Increments 1–4, against an 80% target. There were no other process changes during this period.

Roadmap: the roadmap moved out of individual PM slide decks and into shared planning. By mid-2025, the people in the sessions described it as fully collaborative.

Adoption: a Technical Product Engineer took the format to a second squad, and a third squad picked it up after hearing about it from colleagues. A fourth squad started its own version in 2026 Increment 1.

“Alex introduced our squad to the Product Trio concept, and it’s been instrumental in setting our roadmap and preparing features for development.”
— Kirsten Erich, Technical Product Engineer

Reflection

I would set the rule for a forward-looking agenda in the first session, rather than correcting the drift later.

Have a squad that could use a shared roadmap?

Let me know