Subscriptions: The Good, The Bad And The Overused

[00:00:01.510] - Olivier Destrebecq
Welcome to the Subscription League, a podcast by Purchasely. Listen to what's working in subscription apps. In each episode, we invite leaders of the app industry who are mastering the subscription model for mobile apps. To learn more about subscriptions, head to subscriptionleague.com. Let's get started.

[00:00:19.910] - Speaker 2
Hey there. Today is a special episode because we're continuing the great interview that we had with David Hansson. If you don't know what I'm talking about, then you should stop and go listen to the previous episode and then come back to this one. If you've listened to the previous episode, then welcome, stay here, listen to the rest. Enjoy.

[00:00:38.560] - Olivier Destrebecq
I want to switch gear a little bit because I know Jeff is waiting to ask you questions more around subscription, so I'm going to send him that way.

[00:00:48.520] - Speaker 2
Now let's go deep into our business model around subscription. At Purchasely, we are organising this podcast. We are a growth platform helping subscription apps to grow their business and grow their revenues, basically helping marketers to launch their campaigns without having to republish the app and being able to A/B test in no time. Now let's talk about subscriptions, which is the fastest growing business model on the stores and on the apps. There's one initiative that you launched at 37signals about ONCE. That is an alternative to subscription, and I'm very curious to hear about it.

[00:01:21.860] - David Heinemeier Hansson
Sure. When we got started in 2003, we got started before the SaaS term was even coined. I think SaaS, software as a service was coined in 2007 or something else like that. We had to work very hard to get people to understand that buying software on a subscription can be a really good model in a lot of cases, especially when you're buying service software that needs constant maintenance and improvements and so forth. I love SaaS. I build my fortune off SaaS. I don't have anything against SaaS. I think it is actually a wonderful model for a large number of applications.

[00:01:57.240] - David Heinemeier Hansson
But not all of them. I think, unfortunately, to some degree, SaaS has conquered too well. It has succeeded too far. We have now assumed that everything needs to be a subscription all the time, and it's driving people nuts. It's driving me nuts. One of the applications that I use that is a subscription software—I absolutely hate a subscription software—is Adobe Lightroom. I pay a monthly subscription to Adobe to use Lightroom. You know what? I don't want any updates. In fact, it would be a feature for me if I could buy the version that did not have updates. I just I want my photography workflow to stay sealed.

[00:02:33.210] - David Heinemeier Hansson
Then maybe one day, Adobe comes out with a better version of Lightroom like they used to do with Photoshop 4 and Photoshop 5. I might go, "All right, I'll get Photoshop 6. I'm on Photoshop 4, I skip 5, I now want to get 6." That's no longer an option.

[00:02:48.270] - David Heinemeier Hansson
It's every company doing it. Now I look at how I spend my money and I go like, "Subscription, subscription, subscription, subscription, subscription." Every one of them will go, "It's only $15. How much can $15 be?" Yeah, do you know what? It's not just one $15. It's like 15 times $15, all of it. You go like, there's some subscription fatigue that's building, not just amongst consumers, but also amongst businesses. I think it's 100% justified. There's a lot of applications that should not be subscriptions, that should be one-time sales.

[00:03:18.940] - Olivier Destrebecq
How do you draw the line? Or which one do you think should be subscription and which one do you think should be one-time payments?

[00:03:24.440] - David Heinemeier Hansson
The easiest place to draw the line is whether the app depends on a service. It's integral to the app that there's a service on the back-end. Maybe you're providing backups, maybe you're providing a multi-user experience. There's a bunch of reasons. Basecamp is a great example. Basecamp makes no sense as a single-user app. That's just not why you use it. You use it to collaborate with others. People sign up for Basecamp, and then we need to maintain the computers that provide that Basecamp service, and we need to make sure that it continues to secure and all these other things. Perfect subscription business. That's what we run for 20 years.

[00:03:58.270] - David Heinemeier Hansson
Another great example. If we're going to host all your emails on our servers, and we're going to do the work to continuously update the spam filtering and all the other stuff. That's a service. Adobe Lightroom, I want a piece of software that takes the photos off my camera, put them into this interface, and I can edit them. It doesn't require any service. I don't store my photos in the Adobe cloud. I actually think it's not a very good cloud at all. I prefer to use another service for that Dropbox or just to back it up myself. That's a terrible fit for it.

[00:04:24.870] - David Heinemeier Hansson
I think a lot of types of games, one of single-player experiences. Do you know what? They don't make sense of subscription software either. Multiplayer games do. This is where even within one category, you can find like, all right, great fit for service. If you're playing a multiplayer game, that works well for that. If you're playing a single player game, do you know what? I just want to buy it once. You can try to entice me again with DLCs or upsells or whatnot. I don't know if it's always super clear and cut, but you can certainly see archetypes.

[00:04:52.880] - Speaker 2
What about dating apps? They are the most profitable apps on the stores.

[00:04:55.590] - David Heinemeier Hansson
Maybe as a single app, but as a category, games are way, way, way above that.

[00:04:59.710] - Speaker 2
Apps, not games. I split app and games.

[00:05:02.230] - David Heinemeier Hansson
Yes. I mean, that's one of the reasons Match is running. It's a good question. How would you do a dating app without a service? Maybe you could do one that's based on locality. You're just in local area and you have access. But otherwise, if you're going to see a broader set, someone's got to be storing that data.

[00:05:17.060] - Speaker 2
Because usually the dating apps, what they do is that they offer you only 3 months, 6 months or 12 months. They don't ever, except maybe some few examples, but the bigger ones, they never offer a one-week trial because once you unlock that one week, you get all the value out of it. All the profiles that unlock at the same time. That's making me think that it's not a subscription business model. It's just like they force you into a subscription business model, but it could be a one month.

[00:05:41.340] - David Heinemeier Hansson
I would actually say that some of the problems with dating apps is that it's a subscription model. They're actually incentivised for you not to find a life partner. It is the worst possible outcome for them is if you hook up with a partner. There's a lot of social criticism of what happened to Tinder culture. You can see, you know what? Unfortunately, all the negative aspects of Tinder culture and disposable relationships or so on, they're downstream from the incentives that fit well with that business category, and that's not great.

[00:06:09.790] - David Heinemeier Hansson
Some of it is an uphill battle because we're finding some of those core incentives where it is in many cases, vastly more profitable to operate a subscription business than it is to operate a product business. A product business, you have to keep finding new customers, and you have to sell those customers on what you have to offer. Then maybe occasionally, you get the opportunity to entice existing customers to buy a new version, but it's very lumpy. There's a reason why the financial world have woken up to just what amazing business SaaS software can be because it's so stable. You sign up and you RRR and your MRR, they just look like the perfect curves.

[00:06:47.660] - David Heinemeier Hansson
There's no oscillation. It doesn't go up and down. Again, I'm a beneficiary of that. We've had a very stable, very profitable, very successful business for 20 years doing exactly that. I am not saying that's bad. The reason I'm saying that is in the case of something like Basecamp, I do feel like it's in symbiotic relationship. One of the reasons I say that is, so we've introduced ONCE. once.com is a new line-up products where you buy them once, and you run them yourself.

[00:07:12.890] - David Heinemeier Hansson
Our first product on that is called Campfire. It's essentially an alternative to Slack, which is sold as a subscription software. Slack is $15 a person or something. We sell a competing system to Slack called Campfire for $300 one time. The only way we can do that and still stay in business is that then you have to run it yourself. That's part of the product idea here is that you buy this thing from me, but now it's your problem. I can't have ongoing maintenance cost maintaining your system if you only pay me once.

[00:07:44.170] - David Heinemeier Hansson
I think there are subscription software that occasionally when they get desperate, they try that. They try things like lifetime subscription, which usually means that the business is about to go out of business in 18 months. That's what lifetime subscription means. It's not a good model. You're not incentivised. You're actually incentivised if you sell something once, and you have to maintain it forever, you kind of want the user to die. That would be the best possible outcome is they drop dead the day after they buy your lifetime subscription.

[00:08:08.310] - David Heinemeier Hansson
If you want a symbiotic relationship, you need something where the service continues to offer value over the long term, and it feels good. What we've had with ONCE is we've had some customers go, "All right, this is actually a great deal." Especially in the case of Slack versus Campfire. If you have a thousand users on Slack, it's $15,000 a month. A month. It's crazy. I mean, that's almost $200,000 a year to use a system with just a thousand users on it. In comparison to that, Campfire is $300. One of the problems we've had, and this is one of the general mysteries or fun things about pricing, is pricing could be so far apart that your customers won't believe you.

[00:08:44.520] - David Heinemeier Hansson
When we try to pitch a Slack user or a company that is paying $15,000 a month, they don't believe that there's a working alternative that costs 99% less because it doesn't seem credible. It seems like that can't possibly be. We faced some of that when you start a new product category, just like we did when we started with SaaS over 20 years ago, you have to fight against expectations that customers have of a category. They go like, "This can't be serious. They can't sell this for $300 and still somehow make money and stay in business." Of course, you can. The longest part of the software industry, if you look back 40, 50 years, it was all product-based. It was not subscriptions. Subscription is a very modern thing. I mean, 15 years?

[00:09:27.360] - Speaker 2
Yeah, it got democratised with Netflix. It's pretty recent, yeah.

[00:09:32.780] - David Heinemeier Hansson
It is very recent. Again, I like it.

[00:09:35.270] - Speaker 2
Even on the App Store, I think it's 2017, '16, something like that.

[00:09:38.750] - David Heinemeier Hansson
I understand why people are excited about it. I understand if you run a software business, and you're used to the lumpy revenue of a hits-driven product business, you can look envious at subscription software and go like, "That would be nice if we just could steadily make X dollars a month." But again, it's also, I don't know if it's a strategy of the commons, but it's a strategy of overuse. When then everyone takes everything to subscriptions, eventually there's going to be a consumer backlash. I think we're just in the early phases of that backlash where people go like, "I'm not going to pay a subscription to 25 different things. That's just not reasonable."

[00:10:16.250] - Olivier Destrebecq
Yeah, it definitely started. I'm curious how you price your products, actually, whether it's subscription or product at ONCE. Do you price by geography? Do you have discounts or promotional offers? How do you go about it?

[00:10:29.580] - David Heinemeier Hansson
The first strategy we use when we pick initial pricing is finger in the air. What feels right.

[00:10:35.310] - Olivier Destrebecq
It works pretty well.

[00:10:36.490] - David Heinemeier Hansson
It does work pretty well. It's informed by our own gut and our own sensibilities. Would I pay for this? I mean, yeah, I'm this kind of company, I remember being that kind of company. I'm still that kind of company. I'm incredibly cheap when it comes to these services. We cancel stuff all the time, even if it's small amounts, because you know what? I find an operational beauty in a lean organisation that doesn't squander its resources.

[00:10:58.870] - David Heinemeier Hansson
I feel very in tune with this idea of value. Is this worth it? Is it not worth it? Now, we've tried a lot of things over the years, though, because that might give you a starting point, and then you should experiment. You might find out that there are customers who are very different from what you think, who might pay a lot more, or you might find that there's a bunch of customers who might pay a lot less. We have tried almost everything on the spectrum. We've tried single-price, we've tried multiple tiers, we've tried per-seat, we've tried a lot of different things. One of the things we've tried most recently is international discount. We've tried different pricing in India, Brazil, a handful of other companies with mixed results, I'd say, it wasn't as easy as…

[00:11:38.480] - David Heinemeier Hansson
Sometimes you think like, "Oh, all we have to do is if we make a special offer in India, they're all going to swarm over it." No, I think actually India was a good case where we lowered the price by two-thirds, and we did not increase signups by 3X. If we didn't do that, it didn't net out. There were some increase in signups, but not enough to offset that we lowered the price by two-thirds. That experiment didn't pan out.

[00:12:01.770] - Olivier Destrebecq
Do you try the other way around, raising the price, see if that changed the needle?

[00:12:04.970] - David Heinemeier Hansson
Actually, that one we have tried. We did that with Basecamp. Most famously, we used to have three tiers, 3–4 tiers. When we initially launched 20 years ago, the cheapest one was $20. I think then when we introduced a single price, maybe it was $30, and we went to $99. Flat price. We 3Xed the lowest tier. That one panned out, at least in the medium term.

[00:12:29.180] - David Heinemeier Hansson
We have some reservations about what happened in the very long term, like 3, 4, 5 years in, but we were very rigorous about that rollout, that price increase. We spent six months A/B testing, and we finally got confident enough that it could work. Then after a year, even two years, it was still net positive. We obviously lost some signups, but we more than made up for those signups by the increased revenue. It's very hard, I find, to make generic statements about these things because I'm involved with a handful of other SaaS apps. There's not that much to transfer. You get these ideas, "Oh, here's a best practice." Yeah, I don't know. Maybe for dentists? I don't know if it works for construction workers. The human psychology around pricing is fascinatingly complicated, and we can come up with all these theories about why things are the way they are. But so often we're just shocked when we start experimenting with reality.

[00:13:21.470] - Speaker 2
Yeah, and beyond the numbers, there's even side effects that we wouldn't expect. We had TomTom, the navigation app in a previous episode of the show, and they were sharing with us the fact that they found out that the price point at €40 per month were better, like bringing more money than $20. But the drop of users was so high that they didn't feel confident to set the pricing to 40 because they think, "Hey, we have more people at 20 that could be referring the app, that we could also work later on to increase, upsell or whatever. The drop is too important at €40, even if it's more profitable." It's not only a matter of pricing.

[00:13:58.650] - David Heinemeier Hansson
I think that's actually, maybe that does sound like a pseudo-generic realisation because it's the kind of the realisation we've gotten to, too. That we accepted a substantial drop in signups, which did not pan out in the short term, even medium term as a problem. But if you look over this long term, and you go like, referral cycles and loops and whatever, it's not looking quite as daisy as it once did. This goes against this other meme I've seen online, especially when it comes to B2B SaaS. People go like, "Oh, all you got to do is raise prices." As though companies weren't price sensitive, of course they are.

[00:14:34.550] - David Heinemeier Hansson
This is perhaps the most surprising thing to me in all of these experiments is that the supply demand curve, the pricing versus interest curve, is almost perfect. We tried, I think, six different pricing points at one phase of our experiment, and they all netted out to almost the same thing. You're like, "We'd raise the price a little, signups would go down a little. We'd raise the price more, and signups would go down more." You're like, "Is this on a freaking curve? How is this even mathematically that beautiful? How is the curve that clean?"

[00:15:07.680] - David Heinemeier Hansson
I always just thought, "That's just such a core of human behaviour." Which I'm also surprised about. We have thousands of people sign up for Basecamp every week. It's mostly the same number of thousands. It doesn't bounce around crazily unless we have some huge media event or otherwise. How is it that 3,000 people, week after week, decide to sign up? I think humans are very interesting to study, and we're very much aggregate heard statistics at some level, even though we're all quirky individuals at the micro level. Once you zoom out large enough, the patterns repeat. I think that's also true for testing. The only thing that will tell you the truth is experimentation.

[00:15:48.760] - David Heinemeier Hansson
Now, there's also a deadline for that. We spent six months experimenting, and by the end of the six months, we had experimentation fatigue, and I think so did our prospective customers. They would show up one week, and they'd see one price, and they show up next month, they see a different price. I don't think there's a limit to how much you can experiment. You can't constantly be varying your prices and make a consistent marketing message. I think you need to keep that in mind. At some point, you got to just pick.

[00:16:11.530] - Olivier Destrebecq
It sounds like you guys do a lot of experimenting and probably not just on the prices, but also about feature. I'm curious if there was an A/B test that you guys started, you would have bet the company on the result, and then the test went totally the other way around?

[00:16:26.550] - David Heinemeier Hansson
We don't actually, or we've stopped A/B testing in the product.

[00:16:30.260] - Olivier Destrebecq
Oh, interesting.

[00:16:30.910] - David Heinemeier Hansson
We don't try to guide product development roadmap from A/B test. I found that it's very often very muddied, open to interpretation, and we're better off having a strong cohesive gut instinct as to where the product should go and what features it should have and which features it shouldn't have. Where we've done the vast majority of our experimentation in A/B testing has been pricing and on marketing pages.

[00:16:52.050] - David Heinemeier Hansson
It's funny with the marketing pages, I'll tell you the opposite of that, which was at one point we had a CRM system called Highrise. We still have it, it's accessible to existing customers, but we no longer sell. But at one point we had handed that over to another team outside of our company, who were running it, and they changed the marketing page, the front page, and like, "Oh, yeah, we all liked it better." Like, "Oh, this looks cleaner. It looks really great." Then no one fucking ran the A/B test.

[00:17:15.160] - David Heinemeier Hansson
Six months later, we find out there was a 30% drop in signups. You're like, "Holy mother of…" That was a multimillion-dollar mistake. Multimillion-dollar mistake. It came from, whether we had tested or not tested as an A/B, you should have done that, right? I think. But at the very least, just pay attention when you make consequential changes like that.

[00:17:39.440] - Olivier Destrebecq
Taking 6 months to realise it is a long time.

[00:17:41.370] - David Heinemeier Hansson
That was painful. That was a painful realisation that you should be paying attention to those things when you do changes. It's not obvious at all. We had another, actually for that same product, too, we had another marketing page which had these hero cards with people on them. We found out that Joy was her name. Joy just out-performed everyone. Do you know what? If maybe you could imagine a stereotype here about, do you know what? It's an attractive woman or something. That wasn't it.

[00:18:07.960] - David Heinemeier Hansson
If you had asked me for the five hero cards we had run before, I would not have bet Joy, right? That Joy was going to be the winner. But Joy was the winner. We ran with Joy for quite a long time, and she was like this large hero card. From whatever that segment was, that just resonated just right. I remained a big fan of that as a just insurance policy that you don't screw up your business because these things are quite subtle. They're hard to reason about. Sometimes it don't make sense at all. You're doing these alterations, you're like, "I don't know why this works better. But the data says that it does."

[00:18:38.150] - Olivier Destrebecq
It's not even so much to drive the decision, but more as a gut check afterwards. Like, "Hey, we really think this is better, but are we just running into the wall with it?"

[00:18:46.720] - David Heinemeier Hansson
Yes, I think that is quite important, and you should do that a lot of the times. Not all of the times. I think there are times where you decide, "You know what? We believe in this. We're going to go with this." Even if it looks in the short term like it goes down. Like, for example, this experiment we had with raising the prices a lot, very rigorous. The most rigorous we've ever been about any decision ever in the history of this company. Six months of a careful analysis and cohorts and long runs and statistical significance. We come to a pretty conclusive answer. This is better. We should raise prices.

[00:19:18.190] - David Heinemeier Hansson
Then now you get a couple of years into it, and now it's not so sure. This is what's so great about this stuff is that even when you think you're right, even when you think you have it all figured out, do you know what? The universe is really complicated. There's a lot of variables, you can't control all of them. Sometimes what looks like a slam dunk might not be so on a longer time frame or in a different context. Sometimes things also just change. There's a half life of facts. There's a half life of context of window openings, and suddenly it's gone. What worked five years ago no longer works.

[00:19:50.550] - Speaker 2
Yeah, definitely. We've seen that with some customers rechecking while they A/B test a year after, and it turned out that the good variant wasn't that good any more. Also, there's a lot of things around localisation, as you were saying about that hero card, et cetera. Localising these hero card to different countries or geos is also providing a lot of growth in some cases.

[00:20:13.020] - David Heinemeier Hansson
I think that's where, by the way, we're spoiled in the US because it's such a humongous integrated market. You got 330 million people with the same language, with the same currency, with the same everything. I mean, I've invested in a handful of startups in Copenhagen, I knew it was hard in Europe, but I didn't appreciate just how hard. Like how small the individual markets are and how little they transfer. You do really well in Denmark. Well, that doesn't tell you anything about Sweden. It doesn't tell you anything about Germany.

[00:20:44.540] - Speaker 2
It doesn't transfer. The regulations and it's really hard to sell. Well, US is our first market for Purchasely, even though that we are EU-based, mostly. It's pretty crazy to think about that. Thanks a lot for sharing all these elements. I think that people listening to that show so far have seen that you have pretty singular way of doing things and thinking, et cetera. Just one last thing to ask you about is what should other companies copy from your companies?

[00:21:12.170] - David Heinemeier Hansson
Our scepticism, our first principles approach, because I think as soon as you copy any particular tactic or method, you're going to find that maybe it doesn't work so well for you. Maybe it doesn't work in this context. What you got to copy is a way of thinking more than a set of results, more than a set of outcomes. Because even for us, our outcomes and our results, they're not permanent. As we talked about half life of facts, all these things set in. Developing a scepticism towards, in part, best practices, in part, all of this stuff that is told like, "This is what you have to do to build a business. Oh, just raise your prices." It doesn't always work.

[00:21:48.170] - David Heinemeier Hansson
I think developing a scientific mindset is one part of it. Then you have to keep that part of your brain alive, the scientific mindset, while also accepting that this isn't science. This is not a physics experiment. You can't just run the collider one more time and get a result that you can trust. There's a trillion uncontrolled variables in every single pricing and marketing experiment. You got to accept that [inaudible 00:22:15].

[00:22:15.730] - David Heinemeier Hansson
Some of it is an art. I think this is why it's beautiful, and this is why it appeals to me that you have to use both sides of your brain. You both have to develop some statistical rigours, so you don't make stupid mistakes like changing your homepage and realising 30% loss in signups after 6 months, while also having a confidence saying, "Do you know what? If I just copy the "best practices" off the three plans, and they all have similar pricing, do you know what? I'm going to be a me too run-along fish here, and that might not also work either. I have to have some individuality. I have to trust my own gut instinct some of the time."

[00:22:51.330] - David Heinemeier Hansson
Then, you know what? You weave all of that in and you try your best. Sometimes your best isn't good enough. It's also part of it. It's to accept. I mean, we've launched products in the long history we've been as a company. I thought they were amazing products. Really good. "Oh, man, they're going to love it. They're going to love it." They didn't love it. They hated it. They didn't hate it. Actually, very rarely do you get anyone who hates anything. The vast majority of people, they just don't care. They just don't care.

[00:23:20.180] - David Heinemeier Hansson
You fall into that pit, and you just got to accept, you know what? I can't tell. I can't tell. This is one of the things we mentioned, the books or the writings or whatever. I'll sometimes write this little piece I think like, "Oh man, this is so good. This is going to go viral, whatever." Like crickets. No one cares. Then I write another piece in 10 minutes, and it's just a throwaway thing like it happened when we moved out of the cloud with all our… I just wrote it one morning on a rush, had to go somewhere, push it out there, come back and see that within a couple of days, literally millions of people had seen that. I was like, "I couldn't have guessed that."

[00:23:53.690] - David Heinemeier Hansson
Accepting that reality is going to humble you time and again. No one can predict the future. The best you could do is throw things at reality as often as you can and adjust to the results.

[00:24:05.890] - Olivier Destrebecq
For those of our listeners that want to hear more about those thoughts of yours and follow you online, where can they go? Because you've given us a lot of food for thoughts, and it's always nice to be able to return to those opinions later. I'm curious, where can we find more?

[00:24:19.500] - David Heinemeier Hansson
My personal website is dhh.dk. It has links to my Twitter, my long form writing on HEY World, has link to all my four books, to my products, and everything else I do. dhh.dk is a great place to start.

[00:24:33.620] - Speaker 2
Thanks again, David, for sharing all these. I'm sure that our editors will love, love or hate, maybe, hearing about all these, but for sure, they are not going to be neutral, which is a great thing. Have a great day and thanks again.

[00:24:45.390] - Olivier Destrebecq
Yeah. Thank you, David. That was awesome.

[00:24:46.680] - David Heinemeier Hansson
Thanks for having me on.

[00:24:47.390] - Olivier Destrebecq
On behalf of the Purchasely team, thank you for listening to the Subscription League podcast. If you've enjoyed what you heard, leave us a five-star review on iTunes or other audio platform. To find out more about Purchasely and how we can improve your subscription business, visit purchasely.com. Please hit Subscribe in your podcast player, and don't miss any future episodes. You can also listen to previous episodes at subscriptionleague.com. See you soon.