A launch case is whether this offer should exist, at this price, through this channel, for this buyer. It is not a 4P speech and not a market-entry of a country (unless the product is the entry). In the room, do not name a launch framework. Name the kill tests: demand, attach or adoption, contribution after cost-to-serve, and what has to be true in operations.
A tree that does not tour “marketing”
| Branch | Kill test | Number |
|---|---|---|
| Buyer and job | Who pays, what they stop doing | Budget owner, not “users” |
| Demand / attach | How many will buy in year 1–3 | Attach % to an existing base, or new logos |
| Unit economics | Contribution after variable and CAC | $ / account / year |
| Channel and capability | Can we sell and ship it | Sales cycle, implementation, risk |
Price sits inside unit economics and buyer. Promotion sits inside channel. Product spec sits inside job-to-be-done. You can use those words; you do not need a mnemonic.
If the firm already sells to the same buyer, attach to the installed base is usually the whole case. If they do not, you are closer to market entry.
Worked example: AI coding module for hospitals
Prompt. Chartwell (clinic software in another case) is the wrong client. Here: MedLedger sells revenue-cycle software to US community hospitals. 800 live hospitals. ACV $220k. They built CodePilot, an AI that drafts inpatient coding from the clinical note. List $48k / hospital / year. Sales wants to launch at HIMSS next month. CFO wants a go / no-go.
Buyer. The HIM director does not write the check. CFO / revenue cycle VP does. The job is: fewer outsourced coding FTEs and fewer denials, without a coding accuracy scandal.
Demand is attach, not TAM. Do not size “US hospital AI.” Start with 800 customers. Sales claims 40% attach in 18 months. History of their last add-on (denial analytics): 11% attach in 24 months. Use 18% as a stretch case, 11% as base.
- Base: 800 × 0.11 × $48k = $4.2m ARR at plateau.
- Stretch: 800 × 0.18 × $48k = $6.9m.
Gross margin / cost-to-serve. Model inference $4k / hospital / year. Implementation 40 hours × $150 = $6k, amortized over 3 years = $2k / year. COGS-like $6k. Contribution $42k / hospital / year. Fine, if they sell it.
CAC. A hospital add-on still needs a 4-month security review. Fully loaded AE + SE time: $22k per closed add-on. Payback on contribution: 22 / 42 ≈ 0.5 years. CAC is not the killer.
The killer is accuracy risk and the attach rate. If coding error rate is 2% worse than the hospital’s coders, denials wipe the $48k. A pilot on 20 hospitals with a holdout coding team is the real launch. HIMSS keynote is not.
Capability. MedLedger’s AEs sell RCM suites, not clinical-NLP. They will need a specialist SE. Hiring three before 20-hospital evidence is inventory, not launch.
Recommendation. Do not “launch.” Run a 20-site paid pilot at $24k (half list) with accuracy SLAs. Kill if attach interest <15% of the 80 hospitals offered, or if denial rate worsens. Only then list at $48k. Year-1 ARR target $1m (20 × 48k, or 40 × 24k), not $6.9m. Risk: a pure-play NLP vendor bundles with Epic and the attach never happens — that is why the pilot is paid and time-boxed.
When a launch tree is the wrong tool
Profit of the core product fell. Diagnose profitability before you invent a SKU.
The “new product” is entering a new country. Entry.
The idea is a new growth path among four (same market vs new). Ansoff sequencing first; launch mechanics second.
Org cannot ship anything. If engineering is frozen in a founder war, that is 7S / org, not a GTM plan.
The mistake unique to launch cases
Sizing the universe of hospitals instead of attach to our 800. US community hospitals ~3,400. 3,400 × $48k = $163m “TAM.” You do not have those logos. MedLedger’s reachable market this year is a percentage of 800. TAM theatre is the unique fail.
Second: list price without the budget owner. $48k that hits a HIM cost center that has no budget is not a price. It is a wish.
How to open
“I would treat this as an attach problem on 800 hospitals, not a US TAM. I will test attach using the last add-on’s 11%, contribution after inference cost, and whether accuracy risk makes a full launch irresponsible versus a paid pilot.”
Then put $4.2m plateau next to the sales team’s $6.9–20m. The gap is the case.
Structure a go / no-go, not a brochure
Launch prompts in structure drills should die on attach or unit economics, not on four marketing labels.