How Roku and Output turn failed payments into a subscription growth catalyst

So how do different payment methods impact your churn rates, and what are the strategies that you can implement to optimize those payment methods?

I can start. So, we have so when you go to the checkout at output dot com and to sign up for a subscription, you're presented with a card entry or PayPal. In the past, we offered direct debit with ACH and SEPA, and get into why we no longer do that.

But I think it's it's very important to look at so in the early days, when just getting into the subscription business as the company that was selling perpetual license products prior to subscription, we had a lot of learning to do once we started selling subscription and payments were failing. And we were initially looking at the card category as one single category when in fact there's many payment methods within that category. Right? So there's credit, debit, repaid, reloadable, prepaid, non reloadable, and understanding the churn rates for each of those individually was incredibly important. And seeing the trends between them, and I'll I'll come back to prepaid, but we we have a very offering PayPal was actually demanded by customers when we first launched. That was not offered, and they very much want PayPal. It's become a significant portion of our payment mix.

And PayPal churn rates are lower in than debit. Right? But not as good as credit, but lower than than credit or lower than debit, which is always a win for us given that we have a young user base that are musicians.

And, we have a very high debit card use. And so anytime we're trying to push people to another payment method, it would be, you know, hey. You're on debit. Let's try to get to to PayPal.

Coming back to prepaid, when we first launched churn involuntary churn was at an incredibly high rate.

Through analysis, we realized that much of that was through prepaid cards, which could be viewed as maybe trial fraud or they had no intention of paying, but it was a problem.

And so we analyze the prepaid and on started to understand that there was a difference between the prepaid non reloadable cards and the prepaid reloadable cards. So we quickly said non reloadable makes zero sense for a subscription business. So we change that in our our fraud logic.

However, prepaid reloadable, not desirable. However, there are some regions in the world where that is the predominant payment method, like Italy. Right? So we then created a strategy around prepaid reloadable specific to Italy.

And still today, we allow it in other areas with a let but a more conservative fraud rule. Right? More strict than, say, you would get out of debit and credit, and we look at that customer deeper.

But we so we have varied rates in that understanding those decline rates per payment method within that card category became incredibly important to us and really made a massive impact on our involuntary churn.

You know, we've done a lot of consumer research on prepaid cards and what's interesting is there are two, maybe three hypothesis prepaid card.

One starts with trust.

So if the consumer trusts you, they might not use a prepaid card. If the consumer doesn't trust you and they think you're going to bill them forever and don't have a way to cancel and there's no way you can get out of it, one common strategy for consumers is they'll just use a prepaid card. Because when it runs out of money, that's it.

The other hypothesis is really around the unbanked or underbanked in our country. Sadly, we have a large and growing underbanked population.

And so these folks, you mentioned people who are on unemployment or their unemployment benefits are paid, these folks are great customers, they're good people, but they might be just having financial difficulties or they're not banked well and so they're using prepay card. So if we didn't accept prepaid cards, we would eliminate an entire segment of our customer base that people who like to watch television.

I think that's another reason. The third might be gaming. I don't like that one, but it's true. Consumers might be trying to game. I have trouble thinking about consumers who try to save, you know, eight dollars for a monthly subscription or three dollars for a monthly subscription to game to save three dollars seems like that would be, not a good use of time, but it's certainly a good hypothesis.

It happens with us for sure. Given we have a young user base, we have a thirty day trial at the moment, and they wanna gain that trial. Right? And then you just sign up with a new email address and create that fraud trial, like that trial fraud, over time.

One thing I'll say, I have noticed some of these newer banks like Cash App, Chime, or banks that allow you to create virtual credit cards inside of your bank account, right, is being flagged as prepaid even though it's kind of operates more like a debit card. So that that has been something that I've been watching where we have a a young user base. Right? And they are banking not with the large banks. They're banking with these smaller, newer, line banks, and they're getting their card numbers from that bank. And it's being flagged differently than what it would be if it was from Chase. Right?

Again, that goes right to the trust the hypothesis around trust. If consumers don't trust you, they're not going to give you their actual card number. They'll go to their bank and get a virtual card number. Yes. And some of these banks, some of the smaller ones as well as large ones, they set up a virtual card number, it goes to your regular account. But when the consumer is upset, they just turn off the virtual card number, it stops working.

So then for us, we call that a passive cancellation because they've never contacted us, the payment method just stops working.

How do you, discuss the role of pricing and billing flexibility in reducing churn and improving customer retention?

Well, what I would say is first, Roku offers many different products to consumers. So if you want to have, a lesser cost, you can choose to watch ads. And so, if you want ad free, you can pay more. If you want some ads, you can pay less. If you want a lot of ads, you can almost get it for free. So I think it starts again with the consumer and making sure that there's a price point that the consumer wants to buy at. But as far as billing, the important thing here is the billing date is largely artificial.

It's a random date that happened probably whenever the consumer, at least with Roku's case initially decided to start watching Paramount or Disney or Hulu or something.

And so the billing date itself, that is an artifact of that historical decision, and sticking to the billing date at all costs will cost you money.

So you can certainly retry over a period of time if you happen to discover that a particular card method for a particular bank for a particular consumer works really, really well in the first half of the month, you could change the billing date to the first half of the month. Again, with legal contracts and consumer communications and all that. But there's no reason why you can hang on to this historical artifact of the relationship. You can move it to a different date that works better.

And certainly you can also offer the consumer some features. You can have them, if they're in a full ad free, you can have them downgrade to say an ads version when then they can pay less. So if they're having a tough economic time and they don't want to pay the fifteen dollars a month and they want to pay seven and get some commercials or if they want it to go down to zero and watch full commercial TV, tiers is really important. And then, you know, some people offer the ability to pause a subscription.

It's like a light cancel, right? It's like, Wow, I really like this content.

But I don't want to cancel, but I don't want to keep paying and I don't want to watch ads, so I'll pause that and I'll come back to it later. Right? So I think, you know, again, orienting yourself around the consumer and the consumer's need, is is is where it where those decisions can be made.

Yeah. We offer pause in the cancellation survey, with, you know, moderate success, I would say, of, that someone selects that. It is an option, and it is effective to an extent for us.

On our side, so we offer a historically, we've offered a a monthly plan and an annual plan. We recently tiered that out into three tiers.

And to exactly as Jason says, allowing them to move up and down, of those tiers seamlessly and easily is very important giving them those options. But we see a dramatically, better retention rate in annual versus monthly.

And so we try to push people to the annual. You know, it's in the sign up process for us. We have a a very light questionnaire, and that actually helps us define who this user is. And that actually then we change who is like, which one is first. Is it annual gets highlighted first, or is the monthly option get highlighted first? And those for us, right, we have a lot of professional users, and we have obvious users and some beginners, but we probably at least sit in the hobbyist and professional.

And in the professional, if we can determine based on some of their answers that they are more likely a professional, We've tried to push them to the annual option because they may have the money, They retain better, and we try to push as many people as possible, into that direction.

Think it gets into our next question, which is about best practices for data analytics and insights and how you leverage your data in order to improve retention.

Jason, you want to take that one? Sure. So look, data is critical to pretty much all the decisions that we're making at Roku and all the testing. So it's important to be doing AB testing and that type of thing. But in terms of the recurring payments and the payment failures, I think slicing data as many different ways as you can to get to sort of a knowledge or an insight that you wouldn't have otherwise is really important. So what's an example?

Slicing data by card type, we've already talked about that. How you deal with different card types is important, but also response codes are equally important. So, there are some response codes that are theoretically permanent, like you cannot retry that card or otherwise you'll suffer penalties from the payment networks. So I think again, slicing data by card type, by response, also by issuer.

So every issuer bank out there has their own computer systems that are authorizing these payments. And so each of those authorization systems are operating to the way that that bank wants them to operate. So it's not a uniform thing across all banks around the world that it's going to work a certain way. So when you slice data by certain issuers, you might discover insights.

And then of course, we've talked about this already, but the consumer history with you. So if they're a consumer that has a lot of issues, the way you might engage with that consumer is different than a consumer that has never had an issue or has had one issue the first time ever. So I think, data can lead you to those kinds of insights.

Just as an example, we found a bank that had you know, pretty much extraordinarily high approval rates in the first half of the month and extraordinarily low approval rates in the second half of the month. And I, you know, we have our hypotheses about it, why is that true?

It could be related to the consumer's financial health or financial well-being or it could be related to the bank's own auth policies. But pragmatically, we knew that if we tried a payment in the first half of the month, we're gonna get success or highly likely. If we tried in the second half of the month, it's gonna be highly unlikely to succeed. And since we don't have physical goods, we are digitally selling subscriptions, we have a ton of flexibility and let's just move that billing date to the first half of the month, just to avoid the problem because there's a lot of expense in retrying a payment so you can just avoid that entirely.

We would not have learned that without data and analyzing that data and slicing it in different ways, Because it's not like you're going to find some general learning that's applying universally, it's literally down to the bank, the card type, the actual consumer that you're engaging with. So you have to be able to slice and dice it in those different ways.

Yeah. For for our business, we I would say we spend a decent amount of time looking at the data and utilizing each platform that gives us the data. Right? So you can utilize what you're getting from Recurly is different from what we get when we put Recurly and payment or data altogether and product data into a data warehouse and put Looker on top of it and have business decisions maybe on top of that.

That is one view. And then at the analytics at the payment processor is also very helpful as well. I've I've found that if I analyze I I need to analyze from all inks, right, to see the full picture of where the payment processor, they have no understanding of the subscription in that customer relationship. It's purely transactions.

Right? And that can give you some really interesting insights to to break that down just looking at a transaction level and not necessarily tied to the the life cycle of a subscription.

Where in Looker, our internal Looker, I can really slice the data and pull in product information and try to understand more patterns that way. So I think it's important to use all the tools that you're in your arsenal and understand what the point of view of each tool is, right, and how they play together. That has been successful because I feel like if I looked at only one, it would very much be an incomplete picture.

So in summary and kind of to wrap up, what do you look at for subscription businesses to effectively manage failed payments and involuntary churn?

And what are the the most effective strategies and tactics out there? So we've talked about a lot of different things merchants can do, but what would you say are the most important and most effective?

I would just start with it goes back to the consumer first. The consumer self help is the best way.

And so building a business around and providing the incentives for a consumer to correct the problem is probably one of the best strategies. So when we suspend, entitlement so that a consumer can't watch Paramount or Disney anymore because their payment method has failed, they have a natural incentive to fix the problem if they wanna watch the content on that channel.

Beyond that, all the strategies we've talked about here today using data to discover insights and then actioning those insights. By the way, data for data isn't worth anything, right? It's the question is what action do you take once you learn, some insight from that data and analytics? And those actions can Each one is very small, but then when you add them up over long periods of time, it actually becomes very significant.

Yeah. I would, I would say I need to take it even a step further on the customer, there is no better payment strategy than bringing in the right customer.

Like, starting with you bring in at the top of the funnel, we've had so for a period of time, we had advertising that was not helpful. It was bringing in the the wrong user. We thought we were bringing in decent users. We were not. Payment involuntary churn was going up and up and up.

We changed that strategy, and ever since, it's just involuntary churn. It's going down, down, down, down because we're bringing in the right customer, and that is what's really that's been the biggest solve in payments. I often tell the the team there's always so much, payment trickery that can be done to solve a bad customer. Right.

And it's so true, and that's that's what's been honestly most successful for us. I think when we look at our business, insufficient funds is a is a big problem. It's it's been the number one challenge that we've had forever, and that we've seen improvements through, through fraud rules and fraud testing. We've seen improvement through custom retry logic, directly inside of Recurly and and looking at the retry, where the success and and maybe the challenges are per decline code.

We have modified Dunning length and try to to, Jason's point, get as many pay dates as possible. Right? And understanding we've done it from a different perspective of not moving the bill date, but trying to get done along as many pay dates as possible. And I think that is where understanding your consumer is really important because they the payroll practices in a particular country may be very different. Right? In the US, we get paid frequently.

In other countries, they don't. Maybe it's once a month.

And that is something to very much understand and understand how you can manipulate your done in that way, handle your done communication around that, and understanding that customer and when they are most likely to be paid and update their payment method.

And going back to the the easy credit card update, I think, or the easy card update payment method update is really it's crucial. It needs to be very, very seamless, and painless.