The early attempt should answer a real question about the work.

Early attempts get judged on how good they are, which is the wrong measure and a reliable way to waste them. Their job is to settle something you genuinely do not know: whether the material behaves, whether anyone wants it, whether the hard part is where you assumed. A draft that answers a real question has done its work even if it is bad.

Write down the question before you start, in one sentence. Build the least you can that resolves it. Judge the result against that question rather than how it looks.