What we learned scoring 500 ideas: five patterns that predict outcomes

We've now run five hundred ideas through SaaSValidatr's scoring flow. Some were our own. Most were other founders'. A handful shipped. Most didn't. The patterns in the data are clearer than the popular startup advice would suggest.

The five patterns

• Two or three strong dimensions beat one perfect dimension. Ideas with one 10 and three 4s died. Ideas with four 7s shipped.
• Team split on feasibility = pivot signal. If the team agreed on the problem and split on whether you could build it, the idea was usually right but the scope was wrong.
• High uniqueness + low market viability was always a research project. We stopped letting those through without a separate 'can we afford to learn this' question.
• Ideas submitted by the loudest team member scored highest pre-vote and lowest post-anonymous-vote. This is the founder-bias data, on a large sample.
• Ideas the team voted on within 24 hours moved through the pipeline 3x faster than ones that waited a week.

“The ideas that win are not the ones the team loved. They are the ones the team respected.”

None of this should be surprising. What is surprising is how consistently it holds across industries, company sizes, and founder experience levels. The scoring framework isn't the clever part; the discipline of running it is. The teams that run it weekly accumulate more attempts, get better at calibration, and ship more of the right ideas.