I sell products rather than a subscription service, so it's harder to measure when I've lost a customer.
I'm certain I've lost customers because I haven't had enough frequent updates / new products, and wasn't consistent with my email newsletter. I use a niche toolkit that wasn't ready when 64-bit went mainstream, so I missed being in the first wave of 64-bit ready products and got stuck with a lot of rewrites & retooling. So, just getting the basics right is important!
I did include an optional survey whenever someone uninstalls the trial version of my software. 90% of the feedback was useless (ie "id dint like it"), but occasionally a customer would take the time to write about a bug they'd encountered, or a compatibility problem, so I think it's still worthwhile to include.
Some of the uninstall feedback was a huge red herring - customers kept saying they didn't understand how to configure Photoshop to use my software, or had problems using Photoshop itself, or didn't even have Photoshop, so I thought I'd fix it by developing & including a stripped-back simple photo editor that was preconfigured to work with my software. Some customers loved it & said I'd knocked it out of the park... but it did not move the needle at all and was a colossal waste of time. If I'd looked more closely, I would've noticed those uninstall survey requests had mostly come from Brazil & India, countries I wouldn't have been able to monetize anyway.
I also tag incoming support requests so I can see which areas customers keep having trouble with, so I can measure what might be the biggest holes. For me, customers losing their unlocking serial code was the #1 issue & I solved that almost entirely with an online lost-my-code lookup.