Can you re-run validation? What changes your score and what doesn’t

Yes, you can re-run validation on a project as many times as you like. But before you do, it's worth understanding exactly what moves your score, because it isn't what most people assume.

Validation sessions are versioned. Every time you complete a full validation run, it's saved as its own timestamped session with its own score. Your most recent completed session is always the one that counts for studio access and your current Venture Summary Report, but earlier sessions aren't erased, you can look back at them, and you can compare any two side by side.

The comparison view shows more than score movement. When you compare two sessions, you'll see how your score changed, but you'll also see how your evidence changed alongside it, question by question, assumption by assumption. This is deliberate. A score that moved because you found and uploaded real new customer research looks completely different in this view from a score that moved because you rewrote an answer to sound more confident without anything backing it up.

Here's the part worth being upfront about. If you re-run validation and your answers changed but your evidence level didn't improve anywhere, meaning you didn't add anything new the system could actually see, that session gets a quiet note attached to it: this movement reflects restated answers, not new evidence. It doesn't block you or punish you, but the confidence attached to that session stays capped, and the comparison view makes the pattern visible rather than hiding it.

Why we built it this way. It would be easy to let founders quietly iterate their way to a better looking number by rephrasing weak answers into stronger sounding ones. But a manufactured score isn't useful to you, and it isn't useful to anyone who eventually sees your Venture Summary Report either, an investor, a co-founder, yourself in six months. The whole value of the score comes from it meaning something real. So the system is built to notice the difference between real progress and cosmetic progress, and to show you that difference plainly rather than let you fool yourself.

So when should you actually re-run it? After something real has changed: you talked to more customers and uploaded the notes, you tested pricing and got a signal, you found a channel that's actually working, your team situation changed. Each of those is exactly the kind of new evidence that should move your score, and will.

Re-running validation isn't cheating, it's the whole point of the versioning system. Just make sure what's changed is your evidence, not just your wording.