Hi! I’m a fellow reading-not-rememberer (about 90% accuracy, and it’s been that way basically since I started), so I totally understand the worry 
My first ‘tip’ is just that you probably shouldn’t worry about your accuracy stat too much right now, especially at level 2 - you haven’t really done enough reviews to establish a super rock-solid average, so it’s liable to fluctuate up and down. Also, 87%'s really not bad at all - I think the usual consensus is that if your accuracy’s below 80% that’s where there might be a bit of a problem. Plus, some people’s way of learning just leads to lower accuracy stats - I’m totally happy to bash my head against a wall trying to force stuff into my head after a ton of failures rather than taking a ton of time to memorise stuff during the lessons, so I naturally fail more reviews.
Kun’yomi and on’yomi are messy, because the history of kanji being adopted and so on is messy. There aren’t any hard-and-fast rules, but being aware of the background of things can help identify when it’s likely to be one or the other.
To get to kun’yomi and on’yomi, though: ‘on’ readings are derived from how the Chinese character itself is pronounced in Chinese. ‘Kun’ readings are based on native Japanese words.
For instance, 水 (water) has an on reading of すい, and a kun reading of みず. The native Japanese word for water is みず, and it was presumably みず before they ever had kanji - so that’s why it’s the kun reading. They didn’t have kana back in the day, so just assigning the existing word to the Chinese character was the way forwards.
However, kanji came over alongside the Chinese language, so literate people at the time (and this being ancient times, that was very much a minority) knew, and were reading in, Chinese. That’s why on’yomi are even used in Japanese - a lot of words were borrowed in from Chinese wholesale (with adjustments for pronunciation), which is why other simple words like 本 (ほん) use the on’yomi. And even though 水 is みず in Japanese, Chinese compound words including 水 would probably be pronounced as they were in Chinese, rather than making up a whole new Japanese word.
However, this also gives us clues for when we’re likely to see each kind of reading. Just for one example: words like 食べる (たべる) and 大きい (おおきい) have hiragana attached to them, which implies there’s something about them that would be awkward to represent with a kanji. 食べる is a verb, and 大きい is an adjective, and they can both be conjugated grammatically. Japanese conjugation applies to Japanese words - so we can infer that the word existed beforehand and they just started writing it with kanji. As it’s a Japanese-origin word, we can assume it uses the kun’yomi reading.
Meanwhile, words which are just strings of multiple kanji, like 地下鉄 (ちかてつ, subway), are more likely to be on’yomi, because that’s how Chinese words are formed. Not to say 地下鉄 is strictly how the modern Chinese word for subway is written! But as that’s historically how the words were formed and borrowed in by the literate classes, it’s more likely than not to be the case.
(Important note: please, please remember that as with all language there are a million bazillion exceptions to everything, and at some point you have to bite the bullet and just start remembering words as words rather than collections of readings. Once you start, the readings will kind of fade into the background, but they’ll come to your mind a bit easier too.)
I won’t really get much deeper into it, not least because I’m not an expert, but that’s the kind of stuff you can get out of knowing what kun’yomi and on’yomi are.
Personally, I don’t really think much about them at all. I actually dislike how WK teaches kanji first and vocab later, because kanji stick better in my head when I’ve memorised the words they make up.
Sorry for the rambling post, but again - don’t worry too much this early on, you’re already doing great!