offer-architect: score the offer first
What it does
Section titled “What it does”Diagnoses whether your offer is good enough to spend money behind, before you spend the money. It scores the current offer on the value equation using voice-of-customer evidence rather than opinion, maps the money model (attraction offer, upsell, downsell, continuity), designs guarantees with the refund-exposure math run against your actual published policy, and forecasts unit economics with every formula shown.
The output is the offer scorecard, which is the gate behind the rule that nothing launches on an unscored offer. Plus an upgrade memo: the specific changes ranked by what they would move.
This matters because traffic does not fix a weak offer. It just buys you a faster, more expensive answer.
When to reach for it
Section titled “When to reach for it”- Before any launch. It is a hard gate in new-client-launch.
- “Why isn’t this converting?” when the page and the ads are both competent.
- You are designing a guarantee and want to know what it actually exposes you to.
- You want a money model, not just a product with a price on it.
- A client wants to scale and you suspect the offer is the ceiling.
Where it runs
Section titled “Where it runs”Both. Read-only.
What it needs from you
Section titled “What it needs from you”- The offer: what it is, price, what is included, current guarantee.
- Your real numbers where you have them: margin, current refund rate, AOV or deal value. Every input gets a provenance label, so an owner’s best guess is carried as a believed figure rather than a fact.
- The live policy page URL. It verifies the guarantee against what you actually publish rather than what you say in a brief.
- (Strong input) Voice-of-customer evidence from voc-miner, which is what turns the scoring from opinion into evidence.
What you get back
Section titled “What you get back”- The offer scorecard, scored across the value equation with the evidence behind each axis.
- The money model map, with the gaps named.
- Guarantee designs with refund-exposure math, so you know the downside before you promise it.
- Unit-economics forecasts with the formulas visible, labelled
PRICEDwhere inputs are real andUNPRICEDwhere they rest on a believed number. - The upgrade memo: ranked changes, cheapest meaningful one first.
A worked example
Section titled “A worked example”You: “Score my offer.” ($97 course, 30-day guarantee, 4% refunds)
Scores 6.2 of 10. Strongest axis is dream outcome; weakest is time delay, and the VOC bank backs it: buyers say “I bought three of these and never finished one.” Cheapest upgrade: a 7-day quick-win module that front-loads a result, moving perceived time-to-value without touching price. Guarantee math: extending 30 to 90 days models to +1.8pts refund exposure against a projected conversion lift that clears it, but that lift is
UNPRICED(owner estimate, not measured).
Tips & gotchas
Section titled “Tips & gotchas”- A 6 is common and fixable. The scorecard exists to find the cheapest lever, not to grade you.
UNPRICEDlines are honest, not broken. They mark where a forecast rests on a believed input. Treat them as the things worth measuring next.- It verifies your guarantee against your live policy page. If the brief and the page disagree, the page wins and it flags it.
- Score before you build the page. Rebuilding a page around a fixed offer is cheap; discovering the offer was the problem after a month of spend is not.
Related skills
Section titled “Related skills”A hard gate inside new-client-launch. Feeds page-architect and angle-architect. Strongly improved by voc-miner evidence.