When I collected cards upfront for Planscope, I sent out an email 3 days before their 14 day trial ended. The next email they'd get would be an invoice email when their trial expired, and they'd continue to get these emails each month when they were billed.
When Gmail rolled out tabs, I started getting a lot of people who were unexpectedly billed. My "trial expired emails" were usually lumped into the Promotions tab, which doesn't have as much urgency attached, and thus there was definitely a spike in cancelations after that first invoice (people always see those invoice emails; they sometimes miss the "trial expiring and your card will be billed" emails :-)
When you take cards upfront:
* You'll get much higher trial to paid conversion rates. Mine was about 45%
* You'll get a lot of cancellations after that first bill goes through. If you calculate churn based off of "paid accounts who cancel", it will be artificially inflated. Try tracking it against customers who have paid you at least 2x.
* If you want to make your life a bit easier, automatically refund people who cancel within a given window after their first bill.
* Please don't bill people without sending them an email invoice. It's just wrong.
* Don't stake your business on one-off customers. If someone wants their money back, give it back (within reason, obviously. I'll by default refund the last month.) I've had people come back when they really needed Planscope, refer others, etc. And plus, chargebacks are messy.
I'm no longer collecting cards upfront, and I've finally got MRR growth back on track. I made the mistake of simply changing the signup form / billing code without making heavy modifications to the marketing site and onboarding flow, which caused a huge drop in growth after that change.