Millet (wild yeast/sourdough) fermentation decreasing thickness of the slurry

In order to reproduce a traditional Turkish drink recipe for Boza. Which has a thick consistency, low alcohol content (around 1%), and a slightly acidic sweet flavor.

I'm repeating this two step recipe in batches. First batch is with the starter, subsequent batches are using the "boza" from the previous batch.

Millet Wild Yeast Starter:

1. Grind some millet grains into flour (i.e. using a coffee grinder).
2. Mix 50 g millet flour with 50 ml water.
3. Keep covered with a cheesecloth and wait for it to ferment for 2-3 days.

Boza:

1. Cook 50 g millet grains with 550 ml water.
2. Blend into a slurry which has a thick consistency, let it cool down to room temperature.
3. Blend in 50 g of sugar and start the fermentation either by inoculating with old Boza or the starter (50 g).
4. Let is ferment for 3-4 days.

I'm trying to understand the reason why the following is happening:

The first 2 or 3 batches, I end up having a "Boza" with a very very thin consistency akin to water; after the 3rd or 4th batch or so the final product starts to get the desired thick consistency and further batches after that keep producing the same desired results.

I initially thought, the problem with the first batches could be caused by an enzyme like amylase (from the millet flour used in the starter) breaking down the starch and thus reducing the thickness. To test this idea, I've heated the millet flour to 135C (just shy of browning) and kept it there for more than an hour to deactivate enzymes like amylase that might be present. Still the results were the same, the first batch turned out to be as thin as the first batch from the other iteration.

What's the underlying reason for the first batch turning out this thin?