Involuntary cancels — when the subscriber’s subscription ends due to a payment-related issue — can act as a forceful headwind to growing subscriptions. While payment failures are inevitable, it’s possible to chip away at this problem and have a noticeable impact on retention, boosting customer lifetime value and long-term revenue growth.
This is super useful Reid. Thanks. I’m curious about how these tips are being used in an ‘automated’ way by Substack to help us noobs on here who just want to focus on the publishing. Or are there options for us individually to improve things?
One of the benefits of publishing on Substack and incentives being aligned is that we want to help you retain your paying subscribers just as much as you do. We have effective messaging in place to prevent payment failures, and have some ideas for how we could take it to the next level in the future. It’s an important problem that we think about quite a bit.
Fantastic. Having built a combo Stripe+Mailchimp+User database+Publishing email subscription publishing/payments tool from scratch in 2012, I have a sense of how important this is and how much work it involves. Substack really earns its 10%. Thanks.
Good stuff, Reid. I've run a martial arts gym for almost 20 years now, and this is spot on. Retention is huge in any subscription business! The longer you operate, the more important it becomes as a KPI.
Totally agree — retention is probably *the* most important metric for any consumer subscription business, and its importance only grows over time. But it's often measured the wrong way or misunderstood, leading folks to the wrong conclusions and actions.
I must have reworked the way I measure stuff half a dozen times over the years. Sometimes you sort of get stuck due to "legacy" past measurements, something I struggle with after 17 years of operation (to say the least!).
It reminds me of early days Hulu when our investors and board (all from legacy media companies) just wanted to see churn rate — the retention metric from old-school telecom & cable TV, where there were multi-year contracts with huge early-cancel penalties. Also, all those businesses were in slow-growth, hyper-mature mode, which will lead to low churn rates.
We had to fight tooth-and-nail to get them to focus on cohort-based retention rates, which were directly linked to how long someone stuck around as a paying subscriber and customer lifetime value.
Hello and thanks for this tips! I have a question regarding this point:
For example, when a subscriber approaches their renewal date with an expired credit card, we send an email and dashboard notification asking them to update their payment information (especially helpful for annual subscriptions).
I’m this case how is it possibile to understand when the credit card will expire? I use Stripe and in their API is not possible to get this information? Are there any external tool?
Excellent insight Reid. From my end, I have another perspective to add. Involuntarily paid subs, e.g.
"I never ordered this paid subscription. Please refund. Thx."
I refunded but it was after one month that the person cancelled so I ended up paying the Substack %, i.e. it cost me money. I booked it as a loss and moved on. If that were to happen more often, though, what would be the recommended course of action? Prorated refunds?
In the streaming world, we tried to hold pretty strong on not doing refunds. It's a slippery slope, and there's a good amount of intention in buying a subscription (i.e., it's hard to accidentally pay for a subscription). There was a period (and some merchants still do this) where subscription products used trials somewhat deviously to get people to pay. But even then, the person has to input a credit card, so it shouldn't be shocking to get charged.
Implementing anything related to prorating can get pretty gnarly, pretty quickly. It's usually best to avoid prorating charges and refunds if possible.
There's a whole other world of fraud. In that case, the customer should absolutely be refunded. Due to our Stripe Connect set-up, we mostly have to lean on Stripe to block fraud, at least for now.
Of course, if you don't provide refunds, there's a chance it leads to more chargebacks, which cost even more money. So it's a tricky balance.
Jul 6, 2023·edited Jul 6, 2023Liked by Reid DeRamus
I completely understand. We had a whole dept. dedicated to fraud in my old AAA gaming company and I am well aware that this sub was most likely not accidental. I went with a refund anyway but got the short stick because Substack already cashed their cut so I ended up paying. Lesson learned. If the person then contacts Substack for a refund (after the Substack owner declined a refund), what would have happened, though? It is probably explained somewhere... edit: I found the link. no worries... :) thanks again!
At the risk of speaking on behalf of Substack, I don't think Substack would have any issue refunding the 10% cut as part of any refunds. That said, I'm not sure what the cost structure is for moving these funds, or what's feasible with our current Stripe infrastructure. I imagine this just hasn't been a frequent enough thing to prioritize or bring to our attention. I'll share your feedback internally.
I don't think Substack would provide a refund to the subscriber without the publication's approval (and not even sure that's feasible).
That's very kind of you. I just booked the loss, wouldn't even occur to me to ask you guys for a refund. But it's nice to know that this is something that would be potentially possible. Speaks to your flexibility and customer-oriented approach. I had a nice chat about it with Stripe, too, back then. So all in all, it cost way more than the lost ~$5 if you consider the support time involved.
In general, winback emails/messages to former paid subscribers can be very effective. The winback email (sent to all canceled subs 60 days after their subscription ended) was the best-performing upsell email we sent at Yem, measured by emails sent to paid conversions. The only downside is the send volume is much lower — most (if not all) Substack publications have a much larger group of never-paid free subs than free subs that used to pay. While the winback email was 2x - 4x more effective than other upsell emails, the absolute revenue was much smaller.
Still, it's a good thing to do once you get to a certain amount of canceled paid subscribers. And the winback emails are especially important for involuntary cancels — those subs aren't choosing to cancel and many of them would have never canceled had their payment kept working. Sometimes, these folks don't even know their subscription ended.
One other note. Spotify, Substack, and most subscription services will send a cancel confirmation email after the retry period ends and the subscription ends. In that email, there's usually a button to resubscribe, and that helps get some folks to rejoin right away.
Thanks! Super helpful Reid. i'm thinking about doing personalized emails for all of these though, cuz when I get an auto email from anywhere, substack included, my mind auto shuts off lol
With some of the automated email (like the Boost ones), we want to try to make it as easy for those to go out, so you don't have to worry about going through the actual mechanics of sending upsell, retention, or engagement emails. But they *always* work better if the email feels like it comes from the writer, so customizing the subject, copy, etc. is super important.
Same goes for the one-off email templates, like the subscriber referrals announcement emails.
Let me know if you need any help with filtering for these subscribers and emailing them. The Subscribers tab is your friend - happy to help make sure you have it set-up properly.
This is super useful Reid. Thanks. I’m curious about how these tips are being used in an ‘automated’ way by Substack to help us noobs on here who just want to focus on the publishing. Or are there options for us individually to improve things?
Reposting from Notes here as well, for posterity:
No action is needed by you.
One of the benefits of publishing on Substack and incentives being aligned is that we want to help you retain your paying subscribers just as much as you do. We have effective messaging in place to prevent payment failures, and have some ideas for how we could take it to the next level in the future. It’s an important problem that we think about quite a bit.
Fantastic. Having built a combo Stripe+Mailchimp+User database+Publishing email subscription publishing/payments tool from scratch in 2012, I have a sense of how important this is and how much work it involves. Substack really earns its 10%. Thanks.
Good stuff, Reid. I've run a martial arts gym for almost 20 years now, and this is spot on. Retention is huge in any subscription business! The longer you operate, the more important it becomes as a KPI.
Totally agree — retention is probably *the* most important metric for any consumer subscription business, and its importance only grows over time. But it's often measured the wrong way or misunderstood, leading folks to the wrong conclusions and actions.
Thanks, Andrew! 🙏
I must have reworked the way I measure stuff half a dozen times over the years. Sometimes you sort of get stuck due to "legacy" past measurements, something I struggle with after 17 years of operation (to say the least!).
A million percent.
It reminds me of early days Hulu when our investors and board (all from legacy media companies) just wanted to see churn rate — the retention metric from old-school telecom & cable TV, where there were multi-year contracts with huge early-cancel penalties. Also, all those businesses were in slow-growth, hyper-mature mode, which will lead to low churn rates.
We had to fight tooth-and-nail to get them to focus on cohort-based retention rates, which were directly linked to how long someone stuck around as a paying subscriber and customer lifetime value.
Inertia in a small business is a powerful thing. A company Hulu-sized or bigger is going to have some crazy resistance to change!
Hello and thanks for this tips! I have a question regarding this point:
For example, when a subscriber approaches their renewal date with an expired credit card, we send an email and dashboard notification asking them to update their payment information (especially helpful for annual subscriptions).
I’m this case how is it possibile to understand when the credit card will expire? I use Stripe and in their API is not possible to get this information? Are there any external tool?
Thanks in advance!
Thanks for the kind words & question.
It may depend on your Stripe set-up, but it might be doable with the "exp_month" and "exp_year" objects.
https://stripe.com/docs/api/cards/object#card_object-exp_month
Excellent insight Reid. From my end, I have another perspective to add. Involuntarily paid subs, e.g.
"I never ordered this paid subscription. Please refund. Thx."
I refunded but it was after one month that the person cancelled so I ended up paying the Substack %, i.e. it cost me money. I booked it as a loss and moved on. If that were to happen more often, though, what would be the recommended course of action? Prorated refunds?
In the streaming world, we tried to hold pretty strong on not doing refunds. It's a slippery slope, and there's a good amount of intention in buying a subscription (i.e., it's hard to accidentally pay for a subscription). There was a period (and some merchants still do this) where subscription products used trials somewhat deviously to get people to pay. But even then, the person has to input a credit card, so it shouldn't be shocking to get charged.
Implementing anything related to prorating can get pretty gnarly, pretty quickly. It's usually best to avoid prorating charges and refunds if possible.
There's a whole other world of fraud. In that case, the customer should absolutely be refunded. Due to our Stripe Connect set-up, we mostly have to lean on Stripe to block fraud, at least for now.
Of course, if you don't provide refunds, there's a chance it leads to more chargebacks, which cost even more money. So it's a tricky balance.
Thanks for the note & questions.
I completely understand. We had a whole dept. dedicated to fraud in my old AAA gaming company and I am well aware that this sub was most likely not accidental. I went with a refund anyway but got the short stick because Substack already cashed their cut so I ended up paying. Lesson learned. If the person then contacts Substack for a refund (after the Substack owner declined a refund), what would have happened, though? It is probably explained somewhere... edit: I found the link. no worries... :) thanks again!
At the risk of speaking on behalf of Substack, I don't think Substack would have any issue refunding the 10% cut as part of any refunds. That said, I'm not sure what the cost structure is for moving these funds, or what's feasible with our current Stripe infrastructure. I imagine this just hasn't been a frequent enough thing to prioritize or bring to our attention. I'll share your feedback internally.
I don't think Substack would provide a refund to the subscriber without the publication's approval (and not even sure that's feasible).
That's very kind of you. I just booked the loss, wouldn't even occur to me to ask you guys for a refund. But it's nice to know that this is something that would be potentially possible. Speaks to your flexibility and customer-oriented approach. I had a nice chat about it with Stripe, too, back then. So all in all, it cost way more than the lost ~$5 if you consider the support time involved.
do u ever consider reaching out to people who were involuntary cancels after the fact?
Yes, absolutely!
In general, winback emails/messages to former paid subscribers can be very effective. The winback email (sent to all canceled subs 60 days after their subscription ended) was the best-performing upsell email we sent at Yem, measured by emails sent to paid conversions. The only downside is the send volume is much lower — most (if not all) Substack publications have a much larger group of never-paid free subs than free subs that used to pay. While the winback email was 2x - 4x more effective than other upsell emails, the absolute revenue was much smaller.
Still, it's a good thing to do once you get to a certain amount of canceled paid subscribers. And the winback emails are especially important for involuntary cancels — those subs aren't choosing to cancel and many of them would have never canceled had their payment kept working. Sometimes, these folks don't even know their subscription ended.
One other note. Spotify, Substack, and most subscription services will send a cancel confirmation email after the retry period ends and the subscription ends. In that email, there's usually a button to resubscribe, and that helps get some folks to rejoin right away.
Thanks for dropping the comment, Alex! 🙏
Thanks! Super helpful Reid. i'm thinking about doing personalized emails for all of these though, cuz when I get an auto email from anywhere, substack included, my mind auto shuts off lol
Most definitely.
With some of the automated email (like the Boost ones), we want to try to make it as easy for those to go out, so you don't have to worry about going through the actual mechanics of sending upsell, retention, or engagement emails. But they *always* work better if the email feels like it comes from the writer, so customizing the subject, copy, etc. is super important.
Same goes for the one-off email templates, like the subscriber referrals announcement emails.
Let me know if you need any help with filtering for these subscribers and emailing them. The Subscribers tab is your friend - happy to help make sure you have it set-up properly.