Session 3: Defend the Call
A decision that doesn't move a number is a hobby. In this session you tie your call to the number your boss already watches, see it built and tracked on a real engineering workflow, and write the one page that defends it.
Session 2 was about whether the bet survives your own strategy. This one is about whether it survives your approver.
Enter your cohort code.
Session resources are for Signal to Strategy cohort members. Your code is in the welcome email and pinned in the cohort Slack channel.
Not in the cohort? See the course →
Everything from Session 3.
Working guide
All ten prompts in order, the seven parts of the one-pager, and the receipts table. Start here if you want to just run it.
Download the PDF →Session 3 slides
The full ninety minutes, including the live build and the capstone.
Download the PDF →Launchpad Session 3 pack
The files we ran live: the numbers, the one-pager example, the session 2 recap and the simulated event data.
Download the .zip →Simulated events
The tracked events behind the dashboard. Upload this straight into the weekly check prompt if you don't have an analytics tool connected.
Download the CSV →Three things, not a reading list.
- A defended, ROI-tied product direction
- A weekly check that tells you whether it's working
- A set of prompts you can reuse on every call
Decision to number to one page.
- 1. Build your number. Four questions that turn a decision into a number: what people will do differently, how you'd count it, what your boss already reports, and what would make you stop.
- 2. The live build. Decided bet → spec → prototype → code review → feature flag → four tracked events → dashboard and alert → a weekly check. Every tool is swappable. The stop rule isn't.
- 3. The exec one-pager. Seven parts on one page, with a receipts table that ties every claim to its source.
- 4. Capstone. Pitch it, take pushback from your approver's point of view, and sharpen it.
Everything's in the working guide.
Here's the order.
- Build your number Interviews you, one question at a time
- Cost and return Works out the return in your approver's math, and two other ways
- Spec and tickets Turns the bet into something an engineer can build
- What to track 3 to 5 events, and what each gap means
- The prototype Works in Lovable, Claude, v0 or Bolt
- The dashboard With a connected analytics tool, or from a CSV
- The weekly check Keep going, change course or stop, judged against a rule you wrote first
- Make it run every week A scheduled task, or a calendar reminder
- Draft your one-pager Seven sections and a receipts table
- Find the weak spot The weakest line, and the three hardest questions
Not on Claude?
Every prompt works in ChatGPT, Gemini or any tool that accepts files. Upload the CSV instead of connecting an analytics tool.
Q&A session next week.
Bring whatever breaks when you try this on your own bet.
Questions before then?
Post in the cohort Slack channel.
Momentum Product Co. · Heather Gawel + Babajide Okusanya · Launchpad is fictional. For course use only.