How to choose between two product ideas

Published 1 October 2026 · By Daily Product Idea · Our methodology

Do not compare the totals

An overall score is a summary, and two ideas with the same summary can be very different bets. Compare them dimension by dimension. Two ideas published here both score 59 overall:

• Dropbox for Agents has the strongest problem evidence of the ideas published by 1 October 2026, at 78, but scores 30 for market potential and 25 for differentiation. The problem is real; whether anyone will pay, against free alternatives, is wide open.
• Explain-Before-Merge scores slightly lower on problem evidence, at 74, and higher on market potential, at 42, because its research found people already paying for a manual alternative.

Three more all score 57, each for different reasons:

• LedgerLine is easy to build (80) but less clear about exactly who the customer is (55).
• ReconSheet has a clearer customer (70) and is a little harder to build (72).
• Launch Lane has the clearest customer of the three (72) and is the hardest to build (55).

Neither pair has an obvious winner. What the breakdown gives you is a clear view of what each bet depends on. For how the six dimensions are scored, see how to evaluate a product idea.

Compare how much the scores rest on

Scores from thin evidence move more when you learn something. Check each idea's evidence level as well as its score. Of the three ideas at 57, ReconSheet and Launch Lane are rated Supported, corroborated outside the communities where the problem was found. LedgerLine is rated Indicative, even though it has more problem signals than either. At the same score, the better-supported idea is the safer starting point.

Compare the cost of the first test

Often the deciding question is which idea you can prove wrong fastest and most cheaply. Look at the riskiest assumption in each, and what it takes to test it:

• LedgerLine's first test is an accuracy bake-off: fifty real invoices through a vision model. It needs no customers and can be done in days.
• ReconSheet's needs four to six consecutive weeks of real exports from one consultant, so it takes over a month of calendar time whatever you do.
• Explain-Before-Merge's is a two-week diary study with five to ten developers.
• Dropbox for Agents' is a week of real work with an instrumented prototype across two agents.

If two ideas look equally good, start with the one whose test is faster, or run both first tests in parallel when they are cheap. Their results will usually make the choice for you. See how to validate a product idea for designing those tests.

Compare the fit with you

Two dimensions depend heavily on who is building. Customer access asks whether you can find the first twenty customers. Buildability asks whether a small team can build the first version. Adjust both for yourself. Deal Floor scores 34 for customer access, the lowest of its six. If you already know quote-based tradespeople, that weakness may not apply to you. If you do not, it applies in full.

A short checklist

1. Put the two ideas' dimension scores side by side, not their totals.
2. Name what each idea depends on most.
3. Compare evidence levels: prefer the idea whose score rests on more independent corroboration.
4. Compare the cost and time of each idea's first test.
5. Adjust customer access and buildability for your own contacts and skills.
6. Run the cheapest first test, or both if they are cheap, and let the results decide.

Every researched idea on Daily Product Idea shows its six scores, its evidence level and its first tests, so any two can be compared this way. See how ideas are researched and assessed or browse the researched ideas.