Those are good questions. There are a couple of "extreme" scenarios that might help us figuring out what to do.
Assuming plants eat up all K. Your tap water contains no K, you don't dose anything else that contains K, you dose 20g which (after deduction of substrate, glass thickness, space on top = 80 gal) increases your K level to 30 ppm. Plants consume it, you do a 40% WC, and can dose 20 more grams.
Assuming the K consumption by plants is negligible. Your water contains 7 ppm of K, and you dose 10g of KNO3 which adds ~13g of K. You add 20g of K2SO4. After a week went by, you do a 40% WC, and dose 20 more grams.
In example 1, at the beginning of the second week (second dosing) you are at 30 ppm again. In example 2, your tank would end up at around 75 ppm K after WC and dosing.
In my examples I mentioned some of the unknown variables. If we can't measure them, we can make assumptions. For example, how much of the K is being used by plants? It will be more than 0, and most likely less than the 30ppm that you are adding.
I like to err on the lower side of things. For example, I would assume nutrient uptake of plants would be zero. With a 40% water change, I would only add as much as necessary to bring the 40% up to the target level, so 20g * 0.4 = 8g. This will certainly assure that the tank doesn't run out of K, even though it might be a little less than your target level.
Simple thing I do: I have an EC meter, which measures electrical conductivity. The EC of my tap water is 0.60 mS, while I keep my tanks at ~ 0.95 mS. This doesn't exactly tell me the state of K, but it gives me an overall idea of the macro levels in my tanks. Other use different meters (TDS) or methods.
Sorry if that was too much.