/segmently-unit-economics

UC-3 — Fill the reader's numbers and open the link

Route: gives their funnel numbers (spend, starts, rates, take — a price beside them) (SKILL.md § Router) · Eval: P3 · MCP: uc-3, cases-uc-3 Contents: Turns · Calls · Answer skeleton · Boundary · Stops · Worked example

Turns

• A take the reader states is that product's takeOfTaps (22 % → 0.22) — take = takeOfTaps, never asked, never a fee: fill it and run at once, in this message. A fee is only one the reader states as a fee — their processor's percentage and fixed part (Calls).
• Missing inputs: a chain and a price are not the whole scenario. Ask, one per turn, for what the reader did not give — charges per payer (paymentsCounted), buy-tap → paid (completionRate), volume (daily or monthly budget) — or say plainly that each stays the book's value (monthly 2.86 charges, completion 100 %, $45,000 a month) and label it book.
• Sources, in order of trust: the reader's answers; an ad export (chain.cps = spend ÷ funnel starts, input preparation: print "$<spend> ÷ <starts> starts = $<cps> (your numbers, prepared)" — ask for starts or landing views, never clicks); product analytics (chain.p1/p2/p3, takes); RevenueCat (cases/uc-04-revenuecat.md); a payment provider statement (fee pct/fixed, disputes).

Calls

• No project yet: ue init <slug> --product "<name>~<price>~<cadence>" --platform web — --platform store when they named a store; add the ~<trial> segment only for a trial the reader stated (~none for "no trial"): an unstated trial stays out, is recorded in unstatedTrials, and walks as "<name>: no trial — assumed: not stated — the CLI's init reads it as none" — then the JSON edits in .ue/<slug>.json's scenario.
• JSON edits: chain.*, volume, products[i].price / takeOfTaps / completionRate / paymentsCounted / trialConv, only the deductions the reader has numbers for, label, measuredOn = "<source>, <window>" as the reader named them — when they named neither, "reader-stated, source not given", never a date or a source they did not type — touched. A fee the reader states (their processor's percentage and fixed part) goes in as deductions.fee = { "pct": <pct>, "fixed": <fixed> } with NO preset key — even when the numbers equal a book preset (2.9 % + $0.30). A preset makes walk.lines read the reader's own fee "assumed: the book's <preset> preset" and the link write ~<preset>; the preset is only for a fee the reader did not state (UC-0's fee variant).
• Units in the JSON: rates and shares are fractions (0.029 = 2.9 %) — the link writes percent; touched lists only the chain knobs (cps, p1, p2, p3) you set equal to the book.
• npx -y @segmently/cli ue evaluate .ue/<slug>.json --explain (the project file: only an evaluate of a project prints stage), then npx -y @segmently/cli ue link .ue/<slug>.json --label '<name>'.

Answer skeleton

1. The onboarding screen from this evaluate, when this is the session's first answer (SKILL.md § Session start, point 3): stage.value with its derivedFrom, every plan with its price, ROAS on its valueBasis, the base link, the three offers as printed.
2. The plausibility line, before the KPIs: each number the reader gave placed against the book — chain P10–P90 (cps $1.00–2.50, p1 20–40 %, p2 35–65 %, p3 25–45 %), trial → paid by length, charges (monthly 2.86 / weekly 5.36 / annual 1), dispute lines 0.75 % and 1.5 % — and where it sits: "95 % of buy-taps completing payment is above anything in the book — measured or assumed?"
3. The KPIs, each with its value and its unit (units.kpis), no provenance column: walk.lines is the provenance. Never write "your numbers" on a KPI that depends on a book input. Under the table, words.payback, whole.
4. walk.lines, printed whole, one row each — every input with its class and source.
5. words.breakEvens, printed whole (an unreachable rate says so in words).
6. ONE verdict sentence read off this run.
7. The link: the onboarding screen's base link when this answer printed one (it is this scenario's); else privacy.sentence first when privacy.due, the ue link call, then the link.
8. The boundary: walk.opener, the same walk.lines, then the line of § Boundary.

Boundary

Assembled (SKILL.md § The boundary) from the rendered walk: walk.opener, the same walk.lines, then anything neither the reader nor the book can settle — not testable soon, with the readiness.line days at the volume the run used.

Stops

• A take is never read as a fee, and the reader is never asked which field it is.
• Clicks are never funnel starts: ask for starts or landing views.
• No funnel numbers, only a price → cases/uc-00-launch-card.md.

Worked example

The reader: "cps 2.1, p1 28 %, p2 44 %, p3 31 %, monthly 14.99 take 22 %". No project yet and the message carries their funnel numbers, so this case runs at once; they did not give charges, completion or volume — each stays the book's, labelled book.

npx -y @segmently/cli ue init my-app --product 'Monthly~14.99~1month' --platform web
# edit .ue/my-app.json: chain {2.1, 0.28, 0.44, 0.31}, touched [cps, p1, p2, p3],
# products[0].takeOfTaps 0.22, measuredOn "reader-stated, source not given"
npx -y @segmently/cli ue evaluate .ue/my-app.json --explain

The onboarding screen comes from that evaluate: stage live_no_data (scenario present, chain knobs touched by the reader (cps, p1, p2, p3), no observed inputs), Monthly $14.99, ROAS 0.17 gross, the three offers as printed; its base link is the one below.

Plausibility: cost per start $2.10 sits inside the book's $1.00–2.50; p1 28 %, p2 44 % and p3 31 % sit inside their bands (20–40 %, 35–65 %, 25–45 %).

The payback row is the payback field. Under the table, words.payback — printed whole:

Payback: not covered within 12 months assumes the book's first renewal (monthly 60%) and one steady renewal after it that matches your charges count

The inputs — the evaluate's walk.lines — printed whole, one row each:

• cost per start $2.10, p1 28 %, p2 44 % and p3 31 % — measured: your project file — reader-stated, source not given
• Monthly: $14.99, monthly — measured: your project file (price, cadence)
• Monthly: no trial — assumed: not stated — the CLI's init reads it as none
• Monthly's take 22 % — measured: your project file (the CLI's init split for these products is 30 %, not this)
• Monthly's charges 2.86 — assumed: the book
• Monthly's completion 100 % — assumed: the book
• $45,000 a month — assumed: the book's volume
• the book's first renewal behind payback (monthly 60 %, from words.payback) — assumed: the book

words.breakEvens — printed whole:

• cost per start ≤ $0.36 (from $2.10)
• p1 — no conversion rate at that step clears this scenario on its own (from 28 %)
• p2 — no conversion rate at that step clears this scenario on its own (from 44 %)
• p3 — no conversion rate at that step clears this scenario on its own (from 31 %)

At these numbers the scenario does not clear: −$1.74 per start, ROAS 0.17 gross.

The ue link result says privacy.due true (carries prices, chain, takes), so the privacy sentence comes first, on its own line. It raises no page_shows_gross: the page reads measured= and shows the note in its "Measured on" field. privacy.sentence — printed whole:

This link carries your numbers in plain text — browser history, referrers and analytics can see it. I can give you the JSON file instead.

npx -y @segmently/cli ue link .ue/my-app.json --label 'my numbers'

https://www.segmently.ai/unit-economics?v=2&label=my%20numbers&cps=2.1&p1=28.0&p2=44.0&p3=31.0&vol=45000:budget_month&measured=reader-stated%2C%20source%20not%20given&p=Monthly~14.99~1month~none~22.00~100.0~2.86~&ref=skill

Under it, one line: "measured_missing on the init's link → re-minted with measured=reader-stated, source not given" — the evaluate ran after measuredOn was set and printed none.

The boundary: walk.opener and the same walk.lines — printed whole — then what neither the reader nor the book settles:

These figures are this scenario's arithmetic on the inputs above.

• cost per start $2.10, p1 28 %, p2 44 % and p3 31 % — measured: your project file — reader-stated, source not given
• Monthly: $14.99, monthly — measured: your project file (price, cadence)
• Monthly: no trial — assumed: not stated — the CLI's init reads it as none
• Monthly's take 22 % — measured: your project file (the CLI's init split for these products is 30 %, not this)
• Monthly's charges 2.86 — assumed: the book
• Monthly's completion 100 % — assumed: the book
• $45,000 a month — assumed: the book's volume
• the book's first renewal behind payback (monthly 60 %, from words.payback) — assumed: the book
• the mix of this paywall — not testable soon: the book's volume needs 12 days before it is decidable (readiness.line: "paywall mix decidable after 12 days")