Before you send your product to the first ten people, spend twenty minutes making sure they can reach the part you actually want an opinion on. Nothing wastes early attention like a broken signup, an empty first screen, or no idea what to do once they are in. People burn their goodwill on your setup instead of your product, and you get shallow answers about the wrong problems. Run this list first. It is short on purpose, and every item is something you can check in a private window in under two minutes.
Can a stranger get in without you?
Open your site in a private window and sign up as if you had never seen it. No invite you send by hand, no credentials over DM, no account you made earlier that is already in a good state. If getting in needs you, you are not testing the product — you are testing a demo you are narrating.
Check the email actually arrives, from your production project rather than your local one. This is the single most common thing to be broken on a first deploy, and it is invisible to you because your own account already exists.
Is there something to do once they are in?
An empty state with no first action tells you nothing except that your empty state is empty. Seed enough that the product does its job on the first screen: a sample project, a couple of rows, whatever makes the thing legible without ten minutes of data entry.
If your product genuinely needs their own data before it means anything, say so in a sentence before they start, and make that sentence part of what you are asking about.
Have you removed the bugs you already know about?
Walk the main path yourself, once, slowly. Every bug someone finds is a question you did not get answered — their attention goes on the broken thing rather than on whether the idea is any good. You do not need it polished. You need it not obviously broken in the first thirty seconds.
Have you set the context in one line?
Not a pitch. One sentence so they are judging the thing you built rather than guessing at what it was supposed to be. “It is a way to track expenses for freelancers” is enough. Anything longer starts answering the questions you were about to ask them.
Have you said what you want to know?
“Any feedback welcome” produces “looks great.” Name the thing you are unsure about. If the answers all come back about your logo, that is usually because you did not point anywhere else.
Do you know what you will do with the answers?
Decide in advance what would change your mind. If nobody understands the first screen, do you rewrite it or is that fine? If nobody would pay, do you keep going? Collecting feedback and ignoring it is a slower, more expensive way of not asking.
Write the decision down before the answers come in. It is much harder to be honest about a result once you are looking at it.
Most of this list is one-time setup, and it is the same setup for every product — which is why the Dayzero codebase ships it already done: passwordless auth a stranger can complete, a database that keeps each person’s data to themselves, and a checklist that tracks which of these you can actually evidence. The asking-people part happens on Dayzero itself, where a round on your product earns one on theirs.