Scoring launch markets for a nutraceutical brand: the prompt and the criteria

Picking launch markets is a ranking job. The prompt library entry reads: score and rank Nutraverde's launch markets on acceptability, price point, CAC, and regulatory friction. Four criteria, and they have to be fixed before the chat opens. The tool reads, scores and ranks against those four. Choosing the four columns is the work, and it happens on paper first.

What can ChatGPT actually do for me when I am choosing a market?

It produces a ranked table against criteria you supply. Give it the candidate markets and four named criteria, and you get a score per market per criterion, a rank, and the reasoning behind each cell. That is a first pass on a nutraceutical launch question that otherwise stays unresolved.

What it cannot do is pick the criteria. A brand with a premium price point and thin cash has different columns from a brand chasing volume. Acceptability matters more for an ingredient with a cultural history in one market and none in another. The founder of a nutraceutical brand has that view already. The prompt is where the view gets written down.

So the order matters. Criteria first, on paper, in your own words. Candidate markets second. The chat third. Reverse that order and you get a confident table built on whatever criteria the output supplied for you, which is a table about nothing.

The four criteria come before the chat opens

The four in the Nutraverde entry are acceptability, price point, CAC and regulatory friction. Each one is a business question a founder can defend in front of a partner.

• Acceptability asks whether the category already exists in that market, or whether the first year of spend goes on explaining what the product is.
• Price point asks whether the shelf in that market holds a product positioned where yours is positioned, without a discount that breaks the unit economics.
• CAC asks what it costs to acquire one customer there, given the channels that are actually available to a small team.
• Regulatory friction asks how long the label, the claims and the import paperwork take before the first unit can legally be sold.

Those four are not a template to copy. They are one founder's four. A specialty manufacturer exporting components would pick landed cost, duty, distributor depth and payment terms. An exporter of finished food would pick shelf life, certification, cold chain and retailer concentration. Write your own four and the ranking becomes yours to defend.

Four is a useful ceiling. With eight criteria everything scores medium and nothing ranks. With four, two markets separate on the first read.

How can AI help me grow my business if I still check every line?

It moves the first pass off your own desk, and the checking is where your judgment earns its keep. A founder with no research team can have a scored, ranked, reasoned shortlist before the next partner call. That is the growth: the same operator covering ground that used to need an agency or a hire.

The checking is not optional, and it has a shape. For each cell in the table, ask what the claim rests on. Ask for the source behind every regulatory statement and open it yourself. Where the output gives a range, ask what sits at each end of the range and why. Where two criteria disagree, which happens often with price point and CAC, make the call yourself and write the reason in one line next to the rank.

What you end up holding is a research brief on your own market. The brief is the asset, not the answer. Run it again next quarter when the duty structure moves and the ranking updates. That is the part that compounds.

What does the output look like, and what would an operator overrule?

The output is a table, one row per market, one column per criterion, a total and a rank, with a paragraph of reasoning under each row. Read the reasoning before the rank. The rank is arithmetic on the scores, and the scores are only as honest as the reasoning behind them.

Three things get overruled in practice. Regulatory friction is usually scored too optimistically, because public summaries of import rules read cleaner than the actual process. Acceptability gets inflated where a market has a large diaspora, which is a channel, not a category. And CAC estimates drawn from general benchmarks rarely survive contact with a founder who has bought media in that market.

Overruling is the point. Change the score, keep the criterion, and note why you moved it. Now the table carries your judgment and the reasoning carries the evidence. That document goes to your partners as your view, with the working shown. This is the kind of work I run at A.D.A.P.T. AI Coaching.

How do I pick which markets to score?

What if I already have one market in mind?

Score it against three others anyway. A single-market case has nothing to fail against, so it always reads strong. The other three do not need to be serious candidates. They need to be different enough that your four criteria have to discriminate, which is how you find out whether the criteria work.

Can I trust the regulatory friction score?

Treat it as a question list, not a score. Take every regulatory claim in the table, open the source, and confirm it against the authority for that market. Then price the time. A founder who has done this once has a checklist that makes the next market assessment faster and far harder to get wrong.