Product frameworks in practice

Product frameworks have a bad reputation, and they've earned it. Most teams that adopt RICE end up with confidence theater. Most startup OKRs are a to-do list with a new name. The trouble comes from adopting the ritual without the judgment that makes it work. These pages are my field notes on that judgment.

Where these notes come from

Every framework written about here is one I've run with real consequences attached. I scored and sequenced roadmaps at CaaStle across a ~$50M+ ARR subscription portfolio, where a mis-prioritized quarter had a dollar figure. I've set (and blown, and reset) OKRs at my own two companies, Medzin and WisOwl AI. I've used jobs-to-be-done to reposition products, north star metrics to end executive metric arguments, and cohort analysis to find retention wins hiding inside a blended churn number.

The product frameworks covered here

• Prioritization: how RICE scoring actually behaves under pressure, and where it fails.
• Goal-setting: OKR planning that fits startup reality instead of Google cosplay.
• Clarity of ownership: using RACI to unstick cross-functional launches without drowning in process.
• Strategy: jobs-to-be-done as an interviewing discipline.
• Measurement: choosing a north star metric, running honest A/B tests, and reading retention cohorts without fooling yourself.
• Growth: what product-led growth demands from your product before it produces anything for your pipeline.

Across all of them, frameworks support decisions; people still make them. They make trade-offs visible and arguments productive. The moment a framework's output overrides what you're seeing in user behavior, you've stopped doing product management and started doing paperwork. If your team is somewhere in that transition, these essays or a conversation might help.

Frequently asked questions

Which product framework should a startup adopt first?

A weekly user-feedback loop paired with one honest activation-and-retention dashboard. That isn't a framework with an acronym, which is exactly why it works. RICE and OKRs come later, once there's enough surface area that trade-offs need structure.

Do frameworks matter more or less in the AI era?

More. When you can build ten things fast with AI tooling, choosing which ten matters more than ever, and evaluation discipline (the unglamorous cousin of these frameworks) becomes the core PM skill.

Can you train our team on these frameworks?

I do working sessions: we run your actual backlog through RICE, or draft your actual quarterly OKRs, live. Teams retain what they used on their own problems.

Is there one framework you'd throw away?

None entirely, but I'd retire any framework a team uses to avoid making a decision. The most common offender is endless scoring rounds that relitigate priorities instead of committing to a quarter.

Related pages