For founders
What a founder can know before writing a line of code
A walk through one SeedBasis report, from the idea typed in to the tests that would settle it.
The SeedBasis team · · 3 min read · Markdown version
Most ideas get tested the expensive way. Someone spends three months building, launches, and only then learns that a rival already does it for less, or that the people with the problem never pay to fix it. The research that would have shown it was out there the whole time, scattered across app stores, forums, filings and rivals' pricing pages.
SeedBasis exists to gather that research before the months go in. Here is what that looks like for one idea, start to finish.
It starts with the idea in your own words
You describe the idea the way you would to a friend: the problem, who has it, what you would build. A few short questions follow: where you are with it, what you most need to know, your budget and time, where you would launch first, and, if you like, the rivals you already know about. Nothing about you is assumed beyond what you type.
Your answer to "what do you most need to know" decides what comes first. Ask how hard it is to build, and the Build group leads the report; ask how to reach your first customers, and Marketing does.
About an hour of research
The software then researches the idea through more than 70 live data sources: search volumes for the problem in customers' own words, rivals' prices and traffic, app store reviews, forums, patents, court records, company filings and public web pages. It runs every section on every report: the problem, the market, the competition, whether it can be built, timing and earlier attempts, then marketing, the build, the money, the legal risks and a roadmap.
You can close the page. An email arrives when the report is ready.
What you read first
- A verdict and six scores. Problem, market, competition, feasibility, timing and prior attempts, each scored from 0 to 100 with the sentence behind the score.
- A two-page summary. What the idea depends on, the status of each condition, and what would settle it.
- The ways forward. Up to three ways to go after the idea, with what each would cost to start and how long a first customer would take, and which one the evidence points to under a rule written in code.
- The tests. The checks that would settle the open questions, with what would count as a pass, laid out as a to-do list.
Every finding links to the source it came from and says how far it can be trusted. Where two sources disagree, the report shows both rather than picking one quietly.
What it does not do
It does not tell you whether to build. A report shows where the evidence stands and the options it points to, each with its trade-offs; the decision is yours. It does not talk to your customers either. The tests page is where that part begins, with an interview notes sheet and a short survey form ready to copy for each test that involves asking people.
Working from it
The report is a workspace, not a PDF you file away. You can type into any page, highlight, add notes and pages of your own, and the original research stays in Archives. The agent answers questions about the report, with a link to the finding behind every figure, and makes pages on request: an interview script, outreach messages, a one-pager, a pressure test of the idea. When it is time to show someone, the report downloads as a PDF, a Word copy, slides, a pitch deck or a spreadsheet of the numbers.
Weighing two ideas? Run a report on each and set them side by side on the comparison page: the scores, the cost to start, what winning a customer costs, the time to a first customer and the deciding risk.
Where to start
How it works shows each group of a report, and our method explains how a claim is checked before you see it.