Methodology
Every Dig published here is the output of a structured four-stage evaluation. The stages are deliberately adversarial, because the goal is not to find ideas worth publishing — it’s to find the small fraction worth a small-business operator’s time.
The stages
1. Producer
The producer drafts the candidate. It takes a seed-keyword description of a market or workflow and produces a structured candidate: title, target market, pain evidence, why-now, competitive landscape, leverage angle, estimated time-to-first-revenue.
2. Advocate
The advocate builds the strongest version of the bull case. The advocate doesn’t hedge, doesn’t qualify, and doesn’t consider the bear case. The advocate’s job is to make this idea look as good as it can plausibly be made to look. If the advocate can’t construct a credible case at all, the candidate dies here.
3. Skeptic
The skeptic builds the strongest version of the bear case. The skeptic reads the advocate’s case for context but doesn’t directly refute it — instead, the skeptic surfaces structural and contingent risks the advocate’s framing obscures. Common skeptic findings: data-source unreliability, no defensible moat, unit economics that don’t hold at target scale, no plausible customer acquisition channel, or a value proposition that requires a workflow that doesn’t exist in the target buyer’s business.
4. Judge
The judge reads the advocate’s case and the skeptic’s case and assigns a verdict on a rubric. Verdicts are one of:
- Dig — the skeptic’s concerns are real but the advocate’s case substantively addresses or sidesteps them, and a reasonable small-business operator could test the idea affordably. Digs get published here.
- Hold — the case is structurally interesting but one of the skeptic’s concerns is genuinely unresolved. We archive it. If the underlying market changes, we may revisit it.
- Kill — the skeptic’s case is decisive, or the advocate’s case rested on a premise that fact-checking didn’t support. We archive the verdict; we don’t publish the case.
What the judge labels mean (and what they don’t)
A Dig is not a recommendation to start a business. It is a verdict that the idea is worth the time and small dollar amount required to validate it further with an operator-side discovery process — customer interviews, technical feasibility check, channel test. Most Digs that are validated will still not become real businesses. The role of the verdict is to lower the cost of getting to a no by raising the floor on what we publish.
A Kill is not a claim that the idea has zero merit. It is a verdict that, on the available evidence, the case for spending time on the idea doesn’t survive the skeptic.
Editorial review
The advocate-skeptic-judge process is automated. Once the judge labels a candidate Dig, the candidate enters editorial review by the appropriate desk on the Small Optic Desk. The desk:
- Verifies every numerical claim against its cited source. Claims without verifiable sources are removed or marked [unverified].
- Removes adjectives, hedges, and marketing-tone phrasing the producer or advocate may have introduced.
- Confirms the Dig is meaningfully distinct from sibling Digs on similar markets. Templated rewrites of the same idea with a different industry noun are killed at this stage.
- Adjusts the framing to match the desk’s editorial voice and reader expectations.
Only Digs that pass editorial review are published.
What we don’t do
- We don’t take sponsorship from anyone whose product or market we evaluate.
- We don’t accept submitted Digs from outside parties. (Submitting a seed idea via the contact form is welcome; whether we run a Dig on it is the desk’s call.)
- We don’t take affiliate fees on the products we mention as adjacent.
- We don’t rank or score competitors against each other. We name them when relevant; we don’t recommend one over another.