I keep a pretty robust spreadsheet with all sorts of things (I like a good {or even a bad} spreadsheet). Relevant to this question, I look at a few specific things.
I keep the top 5 times for most all of the official segments (KOMs and Sprints), as well as some of the routes (just the ones that I know I’ll do many times). Zwift only keeps those for 30 days, which doesn’t help much, and you can only see them at the time. I use Strava to get the times for those segments. Yes, I know Strava keeps them all, so why do I need to do it too … because I don’t want to search through all 50 times I’ve done the Fuego Flats Forward Sprint, and I want top 5, not top 3, just because. I keep it all in a sheet, and update as needed. I have some IRL segments in there too.
I find that’s also a good motivator. I wasn’t too excited about this off season, but then I decided I wanted to set PBs (or at least get into my top 5s) on a handful of specific segments (ones that I hadn’t been in the top 5 for a long time). That’s motivating me to ride certain routes, or work harder in certain areas, and so far at least, keeping me Zwifting. I’ve been able to set PBs on the Alpe, Epic, Volcano, among others. Putting a new PB (or top 5) in my sheet makes me feel good about the work, and motivates me to do it again.
I keep track of each month’s miles, time, pace, elevation, etc. I can compare month to month, but I can also compare year over year. How’s this January comparing to Jan '21 or Jan '20. I can see that my average speed is Sept, Oct, and Nov is much lower than last year. But I can also see that it’s because my elevation is significantly more than last year, so the lower pace is to be expected. I haven’t climbed as much in Dec and Jan, so my overall speed is coming back in line.
I don’t know that there’s much scientific basis for this, but I also keep track of my Watts/Heart Rate ratio. How high is my heart rate for a given amount of wattage I’m producing. Seems to me that the better shape I am in, the higher that ratio should go. It’s not exact, as a 30 minute race will give a much higher ratio than a 4 hour ride will. And I’ll get very different results on Alpe de Zwift if I exit at the top vs working some on the way down vs coasting down (heart rate’s still up some, but producing 0 watts, so the overall ratio comes down a lot). So I don’t look at individual events so much as I look at overall averages, and if they are improving; are they improving on similar rides. Again, I don’t know that there’s any science to suggest that’s a valuable metric, but it works for me.
In the end, I think it comes down to what you want to see improvement in, and how you want to track it. Honestly, that’s part of why I give Zwift a bit of a pass on not doing a ton more of that. We all have different goals, and would want different things measured. So I took it on myself to track what I wanted to. That’s evolved some over time, as my goals have changed, and I’ve updated my workbook accordingly. So I’d say, think about your goals, and what you want to track, and then think about the best way to be able to do that. Might be a 3rd party tool, or it might be a spreadsheet, or maybe just a piece of paper tacked to the wall in front of your trainer. Whatever keep you going, do it. Ride On!