Ask HN: When to get customers' billing info?

I'm working on the billing system of a new startup, and I'm trying to think through the plusses and minuses of when to get the credit card info from the customers.

I know that some shops (37signals) get the credit card up front, but say "we'll give you 30 days to cancel, and we won't charge you until the end of that time." This method actually bothers me, as it seems to spin "counting on you not remembering to cancel" as "doing you a favor, so you don't have to get out your credit card to actually sign up down the road." But maybe I'm being overly sensitive on that?

Have you tried multiple approaches? Has anyone done A/B testing on signup flows? I'd love anything you have ... anecdotes, data, whatever.

If it matters, the startup I'm working on at the moment has a freemium model, with a free (but feature-limited) version, and a paid (or multiple tiers of paid) version. The market is freelancers, self-employed people, and, to some extent, small businesses.

I'm curious how it would work if you gave a free month, then billed at the end of every month.

Let's say they sign up on January 1. On February 1st, they receive an email or notice mentioning that their free trial is over, and they'll be billed at the end of the month. If they're not completely satisfied, they can close their account and have February's bill completely refunded. At the end of February, they get prompted for payment information.

Some people will use your service free for two months and then close their account, effectively getting a two month trial period. It's unlikely these people were going to purchase your service anyhow.

The people who were going to pay will pay like normal.

The folks who are on the fence about paying might decide to signup because they've now had a month of knowing that they're 'paying' for this service. Once you're already expecting the bill it doesn't seem like a big deal to pay it.

If you implemented a payment scheme like that, you might make it a policy to never charge people for the month they cancel either. I know it loses a month worth of income, but it feels like you'd be giving departing customers a more positive experience.

For a simple premium model with a 30-day free trial, you've got two reasons not to ask for the CC info until you're going to cut off someone's access to the premium product (you want to maximize the number of prospects, especially as a new entrant). I'd just make sure that you just "locked" someone's premium features/data if their credit card lapsed (if you're doing subscriptions) or if they decline to re-up (if you're allowing month-by-month PAYGO payments) - i.e. don't delete it accidentally when you downgrade an account.

And even if you decide to follow 37signals w/r/t the premium version (ask at the time someone upgrades/purchases, but only but actually charge the card at the end of the first period), there's no reason to ask for a CC for the free version; you're making your sales funnel unnecessarily narrow. You can nudge/bother them to upgrade only when they're already customers.

[I guarantee that no one has a more convoluted billing system than does Dawdle, so I've spent a lot of time dealing with this living hell. While we don't require up-front payment, we use credit cards for authentication as a fraud prevention measure for both creating (free) listings and for our "name your own price" StandingOffers. (I mean, if you're going to name your own price, and we want to charge you automatically when your StandingOffer matches a listing, we kinda sorta need your info up front.) Of course, the authentication charges are 1) expensive and 2) piss off customers, but what are you gonna do?]

I'm the designer for tenderapp.com, and we decided to try putting in credit card information when we need it (after the 30-day trial) instead of up front, like we previously did for Lighthouse. Well, decided may be the wrong word... let's call it a happy accident that it happened that way.

Now that we've had this flow live for a few months, I can't think of a reason a small startup would possibly ask for credit card information up front.

Pros:

- We have one of the lowest barriers to entry to our app. Literally 10 seconds and you have a fully functional install.

- We have lowered the amount of people complaining about us accidentally charging them (i.e. they forget about the 30-day trial and ask for a refund at day 31) to zero.

- We've opened up a lot more channels for follow up. We can contact people weeks/months after their trial expires and see if they'd like an extension or to try again. (In other words, we've removed the negative parts of trying the product -- cancelling)

Cons:

- It's lowered our conversion rate (remember what rate means) and made it extremely unpredictable.

- We don't get to steal as much money from people who forget to cancel their subscriptions.

At the end of the day, I believe in our product and I want as many people to try it as possible. A credit card form isn't going to stop someone from purchasing a product (unless it's really bad), but it sure does stop a lot of people from trying out a product. So I work as hard as I can to remove any roadblocks to people getting their feet wet. From the responses I've had, our customers love it too.

Does your lower conversion rate take into account people who cancel after one month, or ask for a refund because they tripped over the thirty days?