Every business and product are different, so there are no specific numbers you should expect in general. You can sometimes find benchmark data from similar companies, but usually you just compare to your own company over time. E.g. if you do a new marketing campaign and conversion rates go down, then you are probably introducing relatively more bad leads rather a sudden poor performance from sales. Although, in the same situation, an influx of a new kind of customer might just need a different sales strategy without changes to product or marketing.
I have lots of thoughts about the necessity to work and how that can drain a lot of enjoyment out of intellectual pursuits, but I will say that, of profit motivated companies, Honeycomb is an exceptional one to work for filled with wonderful humans who care about each other. I really enjoyed my tenure there, made lifelong friends, and became a better engineer. It's a really special place and I hope they keep that up as they grow.
I can see that it has strong predictive power on company revenue, which the board must be very interested in, but is it really stronger than just a time-series? Perhaps it affords a set of knobs that can be twiddled to try to improve the outcome, but how much of that control is illusory? (After all, most startups fail, even -- especially? -- VC-backed ones.)
Outside of that discussion, going a bit more meta, it feels like the experience left a fairly bitter taste in OP's mouth. I continue to believe that corporations and startups, even friendly ones like Honeycomb, are not pleasant places to work at.