A Day in the Life of a System Thinking Product Manager
Systems thinking is the ability to see the structure that keeps producing the same problem, and to change the structure instead of the problem.
By Jason Abdo
I have seen a lot of commentary in podcasts, substacks, and meetings about being a “systems thinker.” So I pulled a few of the advisors and folks at Product Protégé into a roundtable to pressure test what that actually means in practice, and what it does not mean.
The definition we landed on: systems thinking is the ability to see the structure that keeps producing the same problem, and to change the structure instead of the problem.
What came out of that conversation is the day below. We pulled from teams we coached this year across our clients, with the names and features changed.
Watch how the work of a systems thinker happens, where it gets pressured, and where it stops showing up. Almost none of it happens in the backlog.
7:42 AM | The Slack message before coffee
A VP of Sales pings me.
“Can we squeeze the CSV export into this sprint? Big client asked for it on Friday’s call.”
My instinct at 7:42 AM is to type “let me see what we can do.” That sentence has cost me a ton of capacity at various points in my career.
What I sent instead:
“Happy to run it through intake. Send me the client, the workflow they are stuck in today, and what they do with the file once they have it.”
He replied in four minutes with the workflow.
Turned out the client did not want an export. They wanted a weekly summary in their inbox. A scheduled email. Two days of work instead of three weeks.
Our prioritization gate we teach at Product Protege found the real request in four minutes.
Protégé Tip
Systems thinking gets pressured by early morning “urgency” from stakeholders.
It can quickly stop showing up if the persona you take on is “order taker.
Before anything touches the sprint, run the request through Stage 2 of Product Protégé’s 8 Stage Prioritization Flow, the strategic filter that asks whether a request ladders to how we are making decisions for our roadmap (trade off, intent, risks, success metrics/signals)
8:30 AM | Refinement stalls for twenty five minutes
The team is arguing about acceptance criteria on a permissions epic.
Edge case after edge case.
What happens when an admin deactivates a user mid session.
What happens on shared devices.
Good questions, all of them, and none of them have an answer because nobody in the room knows what the epic is for.
Twenty five minutes in, I stopped it.
“Hey guys, let us take a step back for a second. Let me remind us what success looks like here. The vision is that an admin never calls support to manage their own team. The pillar we picked to get there this year is admin trust and control, and the tradeoff we made was pushing all the self serve billing work into next year. If we get this right, the KPI that moves is admin tasks completed without a support contact. So lets walk through how this epic connects to that.”
Silence.
Then the lead engineer said, “I think this came out of an audit finding last year.”
We killed it and logged it in our icebox (not our backlog) so it stops resurfacing.
When a story gets debated to death, the story is rarely the problem.
Protégé Tip
Systems thinking gets pressured when a team tries to settle strategic uncertainty at the backlog layer. . Run a five minute trace up Product Protégé’s Product Empowerment Pyramid before you rewrite a single acceptance criterion. Story, epic, roadmap, strategy pillar, vision. When you cannot say that chain in a narrative, stop refining and go review or even fix the level above.
10:00 AM | Discovery on a payments flow nobody has built before
So we are going after a new market segment in payment apps.
At this point we are the first ones shipping into this market with this approach, more 0 to 1 than 1 to 2.
Engineering has scoped a build that runs 14 weeks and UX already has the full flow in Figma.
I asked one question. How confident are we that this is how they pay today?
We had three sales anecdotes. That was it.
That puts us in the complex domain. The unknown unknowns, where cause and effect only make sense in hindsight.
A 14 week commitment in the complex domain is a guess with a delivery date on it.
So we probed. Five customer calls, a clickable prototype, and a fake door on the pricing page. One week of learning ahead of a 14 week commitment.
The fake door told us the segment wanted it. The customer calls told us they wanted monthly billing instead of per transaction, which broke the revenue model sitting inside the original scope. Better to learn that in week one.
Protégé Tip
Systems thinking gets pressured by team excitement and delivery momentum. It stops showing up when you treat an emergent, unknown unknown problem as a straight execution problem.
Sort the domain (clear, complicated, complex, or chaotic) before you commit capacity, using Product Protégé’s Sense-Making Framework. Obvious and repeatable work goes straight to a user story. Complex and unknown work gets at least a one week probe with a defined learning goal. Sort first, then pick your method.
11:15 AM | The 1:1 where I almost ruined it
An engineer opened with a data model problem.
I think I knew the answer.
I stopped my Advice Monster and asked,
“What is the real challenge here for you?”
He talked for six minutes.
The data model was fine.
The real challenge was that two other co creator teams write to the same table and nobody owns the schema.
My answer would have solved a problem he did not have.
Then I asked, “And what else?”
He gave me the actual blocker.
He had asked the other team twice and gotten no response. That is an organizational structure problem, and the fix is pretty straightforward.
Protégé Tip
Systems thinking gets pressured by your own ego. It stops showing up when you let the Advice Monster solve technical details while the organizational friction underneath goes undiagnosed. Use two of Product Protégé’s Seven Core Questions. “What is the real challenge here for you?” and “And what else?”
Ask the second question three times before you offer anything that sounds like a solution.
1:30 PM | Design review, and the slow leak
Design walks through the onboarding flow.
Beautiful work.
Then a stakeholder says,
“While we are in here, could we also handle the multi account switcher?”
Scope creep usually arrives as a favor.
I pulled up the Scope-Out section of the epic and read it out loud.
Multi account switching was already listed there with the reason. It depends on the identity refactor that lands in Q1.
Nobody argued. The Scope-Out is a perimeter published before anyone gets attached to the idea.
Protégé Tip
Systems thinking gets pressured by polite cross functional requests. It stops showing up when your scope boundaries live only in your head. Build epics with Product Protégé’s 14 Point Epic Spec and write the Scope-Out section before you write any user stories. Give every excluded item a reason and a revisit date. An exclusion with a date is a promise. An exclusion without one can easily come back in three weeks.
2:45 PM | I held a release
Launch is six days out. Code is done. QA is nearly done.
The Metrics Plan is empty.
We have no event on the primary action, no cohort definition, and no dashboard.
If we shipped it, we would know the servers stayed up and nothing else.
Six weeks later somebody asks if it worked and nobody in the room has a number.
I moved the launch four days.
Instrumentation went in, the dashboard went up, and the King KPI got a baseline number so we have something to compare against.
Four days of delay bought us the ability to learn.
Shipping blind costs a full quarter, because the next roadmap gets built on a guess.
Protégé Tip
Systems thinking gets pressured by shipping deadlines. It stops showing up when the team’s definition of done ends at code deployment. In the Elevate phase, shipping is the starting line. Add one rule to your definition of done that you never bend. No production deploy without a completed Metrics Plan, meaning event names, cohort boundaries, dashboard link, and a baseline King KPI.
4:15 PM | Publishing a miss
Last month we shipped a feature I championed.
Projected a nine percent lift in repeat usage. We got two percent.
The Value Update Report went out today with the two percent in the headline, the hypothesis we got wrong, and what we are doing about it.
I hated writing that report.
I sent it anyway. Once your organization watches you publish a miss without political spin, your projections carry weight because everyone knows you would have told them if it went sideways.
Protégé Tip
Systems thinking gets pressured by your own reputation anxiety. It stops showing up when you bury a weak result or spin a metric to protect a project. The Elevate phase asks you to close the feedback loop in public. Publish the Value Update Report on a fixed schedule, in the same template, whatever the number says. The consistency is what builds trust with your team and your executives.
5:30 PM | The word that ends the day
Last meeting.
A director says, “Do we have the resources to do both?”
I said, “We have two teams and one of them is on the identity refactor until March.”
Small correction, and I make it every time. The word “resources” tells a room that people are interchangeable units of throughput.
Language runs ahead of behavior.
Protégé Tip
Systems thinking gets pressured by corporate jargon. It stops showing up when your own vocabulary frames human partners as industrial throughput. Audit your language for thirty days. Swap resources for co creators, backlog for prioritized options, and done for measured. This is part of culture.
What actually made today different
If you look back, I wrote almost no requirements. I attended many meetings, sent maybe fifteen messages, and shipped zero code.
Here is what I did do.
Moved a request from an inbox into an intake gate.
Traced a story up the Product Empowerment Pyramid and killed it.
Sorted a problem into the right complexity domain before spending money on it.
Asked two questions and stayed quiet.
Evaluated two investments through the Dual Value Lens.
Held a launch for instrumentation.
Published a miss.
Fixed one word.
None of those are features.
That is the job, and it maps directly to the E3 Product Mindset.
Envision the intent so the work has somewhere to ladder.
Empower the structure so decisions happen without you.
Elevate the feedback so the organization learns something it can spend.
None of this needed a new process, a bigger team, or permission from anyone. It needed eight moments where I stopped and looked one level up before I answered. I’m curious, how does system thinking show up in your world?
Until next time,
Jason @
First published in Protege Pulse, the Product Protégé newsletter. Read it on Substack.
