Skills for PMs

Every real-world-data product runs through the same PM work: read the field, check that the data can answer the question, and stay close to what customers need. At Optum, three of those steps stalled for weeks, every time, and more experience didn't help. The workflow itself was the problem, so I built a toolkit that runs each one the same way every time.

Know your product inside and out.

Spend less time on templates, documentation, and summarizing research from multiple sources. Instead, organize it quickly and use your PM skills to prioritize, check feasibility, match customer cases, and frame your pitch.

The full path. The business case is in Define, for the big bets that earn one.

The toolkit The skills inside each phase

Same phases as the path above, opened into lanes. Each card is one skill: the problem it takes off your plate, and what you actually get back. Hover a skill to read what it does.

Plus a shared knowledge base: 14 frameworks and a decision ledger the skills draw on.

Read across for the journey. Read down a lane for the skills inside that phase.

7
Ideate
listen to the market, then decide what to build
voc-synthesis repo · 5-skill pipeline
Customer feedback usually dies in a spreadsheet nobody keeps up. This is a small repo of its own that runs the whole chain for you: clean the transcripts, pull and rank the themes, check for bias, and write a weekly digest. You end up with a living read on what customers actually want.
healthcare-ci-research
Point it at your competitors and it comes back with a cited rundown, their earnings, launches, FDA moves, partnerships, so you're not making roadmap bets without knowing what the rest of the field is doing.
positioning-teardown
Pick a competitor and it breaks down how they actually sell themselves, sorts the real from the hype, and shows you where the open space is. You get a clean report you can hand straight to a stakeholder.
opportunity-solution-tree
It's easy to jump straight to features. This walks you down a Teresa Torres tree, from the outcome you want, to the opportunities, to solutions, to the assumptions you'd need to test, so what you build still ladders back to a real need.
ideate
Got a fuzzy idea you can't quite put into words yet? It opens things up, then narrows back down, and you come out with a few concepts you can actually do something with.
concept-gen
Before you commit to building anything, it spins up a few visual mockups from your real project data and puts them in one tabbed file, so you can compare directions side by side.
pm-advisor
Bring it a decision you're stuck on. It picks the framework that actually fits the situation, strategy, evals, roadmaps, AI products, and gives you a clear recommendation you can act on.
→
5 + 6
Define
make sure the data can answer it, then rank and sequence
data-feasibility
A client asks a research question, and the first thing you need to know is whether your data can even answer it. Give it the question and it tells you which assets cover it, where the gaps are, and a simple green / yellow / red call, plus the effort and confidence numbers you'll need for RICE later.
evidence-reviewer
Works out what it would actually take to prove the case, the endpoints, the right comparator, whether the study design holds up, where bias could creep in, and how high the bar is for this kind of decision. Then it pokes holes in all of it.
dua-prescreen
Before you take a data use case to Legal, this gives you a fast read on the HIPAA and contract risk, what's a non-starter, and what's worth bringing to the real legal conversation. It's a heads-up, not a sign-off.
prioritization-advisor
Most teams reach for RICE every time out of habit. This looks at what you're actually ranking and picks the framework that fits, RICE, ICE, Kano, WSJF, value vs effort, then runs it, so you can defend the order when someone pushes back.
roadmap-planning
Takes your prioritized bets and turns them into a Now / Next / Later roadmap built around outcomes, sequenced by what depends on what, and framed for leadership without backing you into date promises you can't keep.
★ Make the case · big bets only · the business-case skill runs the panel below
business-case orchestrator
This is the one that pulls it all together. It runs the four checks below, then writes the whole thing up as a board-ready case with the narrative and deck. Anything that can't hold up under scrutiny doesn't make it to the slide.
market-sizer
Sizes the opportunity, and keeps you honest about it, separating the big TAM number from what you could realistically win, so the figure on the slide survives the first hard question.
financial-framer
Lays out the money side, revenue, costs, margin, so the economics of the bet are right there on the table instead of hand-waved.
right-to-win-skeptic
Plays the skeptic in the room. It argues why a competitor could just copy you and where your case is softest, so you've already handled the hard objections before the board gets to raise them.
scr-narrative
Nails down the story, situation, complication, resolution, and the so-what, before anyone touches a slide. A small critique panel pushes on it from a few angles first.
deck
Turns the approved story into actual board slides. It won't start building until the narrative is signed off, because making slides before the argument is right is how decks go sideways.
build
→
1 + KPIs
Maintenance
keep an eye on what shipped, feed the next round
KPI monitoring activity, not a packaged skill
Keeps an eye on the metrics after launch, so a dip or a win shows up on its own, instead of when you're scrambling to pull numbers the night before the quarterly review.
voc-synthesis repo · runs again
The same listening pipeline from earlier, pointed at the live product this time. So when you start the next round of ideas, you're working from what people actually said after using it.
2
Always on
there for any step, whenever you need them
rage-research
For the questions one search won't answer. It comes at a topic from a bunch of angles at once and pulls everything together into something you can actually use.
data-scientist
Your technical second opinion on the numbers. It looks over the data pipeline, the model, the stats, and the charts, and tells you whether the analysis actually holds up, in plain business terms.

↻ Maintenance feeds the next Ideate. · The build between Define and Maintenance hands off to the data team (data science, clinical informatics, engineering, then launch and GTM) and carries no PM skills.

What this is, and what it isn't. Every output is a draft for you to sharpen, not a decision it makes for you. The toolkit gets you through the path faster, from the competitive read to the feasibility call to the board case, but you still own the call, and the data, clinical, and engineering teams still build the product.
How the toolkit clears each chokepoint
Competitive read
Chokepoint 01
healthcare-ci-research positioning-teardown
A cited competitive read and a white-space map. The kind of thing a team spent about five weeks on came back in roughly thirty minutes, once, run on public sources. It goes deep, though it won't replace a paid analyst report.
Data feasibility
Chokepoint 02
data-feasibility evidence-reviewer
A simple green / yellow / red feasibility call against the data estate, the RICE inputs, and the evidence bar a client's study has to clear. Same check, same way, every time.
Voice of customer
Chokepoint 03
voc-synthesis
Customer signal turned into themes and demand automatically, so it isn't riding on someone remembering to keep a spreadsheet up to date.