How To Land A Pricing Change
Two case studies and four principles to soften the blow for existing customers
Hi there, it’s Adam. I started this newsletter to provide a no-bullshit, guided approach to solving some of the hardest problems for people and companies. That includes Growth, Product, company building and parenting while working. Subscribe and never miss an issue. If you’re a parent or parenting-curious I’ve got a podcast on fatherhood and startups - check out Startup Dad. Questions you’d like to see me answer? Ask them here.
Q: When you change your pricing, how do you best soften the blow for existing customers?
With the push for improved unit economics amongst venture-backed startups, larger public companies and everyone in between I’ve received a lot of inquiries from people looking to navigate price changes. Over the last few years the price of software has risen more than 2x the rate of inflation and when the AI subsidies go away you can expect it’ll rise again.
In this newsletter article I’ll cover:
The various monetization models and how they map to company types
Two case studies: a price change gone poorly and a price change gone well
How to soften the blow for existing customers
Monetization Models and Company Types
Whether you know it or not, your monetization model exists in an intricate balance between your product value proposition, your market opportunity, and your distribution channels. Brian Balfour called this the Four Fits Framework and it remains the absolute best way to think through the interconnectedness of your growth levers.
There really aren’t that many monetization models out there, so let’s parade through them quickly.
Transactions
I recently got back from a Spring Break trip with my kids and ahead of that trip we needed some items—a few new bathing suits, some swimming shirts, and sunscreen, for example. I went to an actual, physical store (remember those?!) and bought them. They handed me the items and I left the store. Maybe I’ll go back, maybe I won’t. I paid them one time—I didn’t subscribe to the swimsuit of the month or a recurring sunscreen order (shocking, I know). This was a simple, consumer transaction between me and the store.
If you’re an elder millennial or a Gen Xer you’ll also remember that this was how software used to be sold to consumers and businesses. You’d buy these physical items called floppy disks or CDs that had the software on it.
And when you bought the floppy or CD you owned it. Some brave souls are trying to bring this back in different forms.
Subscriptions (Consumer)
And as fun as that was, I’d just be an angry old man yelling at clouds if I didn’t acknowledge that a high-speed internet opened up a whole world of new ways to purchase (and access) software.
If you are a modern consumer of entertainment you’re familiar with the consumer subscription model. This is how Netflix, Spotify, Disney, Hulu, Amazon Prime, Peacock, Paramount Plus, Apple TV (and Music) and most Direct-to-Consumer businesses 😓😓😓charge their customers. Some higher-priced consumer goods like cars, fitness equipment and even mattresses will take a hybrid transactional and subscription approach.
Subscriptions (B2B)
In 1999 Salesforce launched as the first software-as-a-service (SaaS) company where you accessed the software via the internet and didn’t install it off of disks. Unfortunately for the rest of us they haven’t changed their UX since then (I kid, I kid!) but they did revolutionize an entire industry with this monetization model. Now much of your business software is delivered as part of a seat-based subscription.
Usage-based
Within SaaS we’re now seeing the rise of usage-based pricing. My friend Elena has written about this a lot. This is especially popular amongst companies that have a strong product-led growth motion because cost of entry is very low (i.e. zero) and grows as you get more value out of the product. You’ll see this across all parts of the ecosystem—foundational infrastructure companies like Amazon AWS, Snowflake, Cloudflare, Google Cloud and Azure; tooling like Stripe, Square, Plaid, Auth0, etc.; and applications that sit on all of this like Slack, Hubspot, and more.
Even the example above, Figma, has a quasi-usage-based model when you look at going from Free to Paid (“3 collaborative design files” -> “Unlimited Figma files”).
Advertising
You thought maybe I forgot about the monetization model that the internet was built on? The thing that keeps most of the internet free and accessible? Of course not!
Advertising is the primary model that powers the consumer internet. Eventually, it comes for us all. Nothing is safe from the shadowy grasp of advertising. Even if you have a different monetization model… if you reach enough scale, you’ll introduce ads.
This makes sense. If we don’t want to open up our wallets to pay for content then we have to stare at a few banner ads because someone has to pay for it!
The largest players on the consumer internet all leverage (at least partially) an advertising-based business model.
Here are the Top 5 most popular multi-platform web properties in the U.S. based on number of unique visitors (millions) (link)
All of the remaining monetization models are just variants or permutations of the above five. If you think I’ve left out a critical one that can’t be explained by those five please yell at me in the comments.
Case Study 1: Alienate Your Customers, Erode Trust, and Make The Entire Internet Angry At You - What We Did Wrong at Patreon
It was the end of 2017 and if you were a Patreon employee you were looking forward to some much-needed rest at the holidays. Having recently rolled out new branding, positioning and products for the company and now experiencing accelerated growth we were riding high. It would all come crashing down on December 7th because of a botched pricing change.
What follows is the breakdown of the pricing model from 2017, our attempt to change it, and our myriad of mistakes. I’m also simplifying a very complex system here to make it easier to understand so please don’t yell about me on the internet because I’m purposely leaving out a few details.
Here we go…
In 2017, Patreon was a B2B2C membership platform and we had a monetization model similar to Shopify (with a few, slight differences). One difference was that we built online software that was free to use but charged creators who used it a percentage of their platform earnings. At the time that was a flat 5%. So for simplicity if a creator made $100 in a given month on the platform, we would take $5 as a platform fee. If you made $0 on the platform, you were not charged a fee.
There was an additional cost to creators who made money on the platform, but Patreon didn’t extract monetary value from that cost: payment processing or transaction fees. The simplest way to think of this is the fees that credit card companies (and Paypal, et al) charge for a transaction. You can see this spelled out rather clearly on a payment processor’s pricing page – take a look at Stripe’s page.
The Patreon platform works when a fan of a given creator initiates a monthly pledge (a membership subscription) to that creator. For example, you start paying Taylor Swift $5 per month on the Patreon platform. Okay, that seems straightforward – the payment processing fee would be 2.9% + $0.30 on that $5, or $0.445 (rounding down for simplicity to $0.44). If that was the only monthly transaction that happened then everything would be quite straightforward. We’d charge the creator $0.44 for the payment processing fee then we’d charge them ($5 - $0.44) x 5% for the platform fee. That amount would be $0.228 (let’s round to $0.22 for simplicity). The total fees that the creator would pay were $0.44 + 0.22 or $0.66 and they’d pocket $4.34 or 86.8% of the pledge.
Sounds simple! Still with me? Well then let’s complicate it some more.
Most Patrons pay money to more than one creator each month. Using our Taylor Swift example, let’s say I pay $5 to Taylor Swift and $1 to The Dollop podcast because I love them so much (I have eclectic taste!). Or further still, let’s say I pay 25 different creators $1 per month because I’m a serious patron of the arts. If that appeared as 25 separate charges, each creator would eat a 2.9% + $0.30 charge on a $1 transaction. Well that sucks. But don’t worry, when we started Patreon we decided that our goal would be to help creators retain as much of their pledge value as possible. So behind the scenes we actually aggregated all of the payments from an individual patron and charged them one, monthly payment on the 1st of the month. So instead of 2.9% + $0.30 on 25 separate, $1 charges we’d charge you $25 and the 2.9% + $0.30 would just be incurred once.
Still with me? Great, let’s complicate this even more.
Because you’re now only processing one payment you have one processing fee. But you’re paying 25 different creators… so… who pays the processing fee? Well, behind the scenes we were both aggregating payments and then disaggregating them to charge fees. This was even more complicated because it wasn’t like a patron paid 25 creators $1/each. More likely they had a few $1 payments, a few $5 payments, a $10 payment, maybe a $50 payment… you get the idea. So we had to disaggregate the fees and charge them in a weighted-average way to the different creators. No creator paid exactly 2.9% + $0.30 as a result – it was always less. In this way we were able to save creators a lot of money on processing fees.
When we were a smaller company, processing $1M, then $10M, then $100M in payments this all worked out fine. But eventually, some cracks appeared in the foundation:
Unpredictability. Because of the monthly aggregation and disaggregation the processing fees paid by a creator would fluctuate every month. They wouldn’t jump around wildly, but they were different and difficult to predict.
Scale. As we grew rapidly, processing aggregated payments across millions of patrons on the first of every month became a very resource-intensive process. Imagine recreating a Black Friday or Cyber Monday 12x per year.
Membership Model. Most membership models are predicated on something called “monthly anniversary billing.” This means that you get charged the same day of the month, every month. For example, if I buy my subscription on April 22nd I then get charged every month on the 22nd (or as close to it as possible).. This doesn’t work in an aggregated payment processing system because all charges happened on the same day.
Okay, before you yell all the solutions at me in the comments just remember that Patreon was a small-ish company in 2017 with fewer than 100 employees, no payments team, and razor thin margins.
The first plausible solution that we settled on to the growing problems above was to move to a membership model with anniversary billing and therefore disaggregate payments. The problem with this was that creators' payment processing fees would increase… a lot.
Our #1 core behavior was to “put creators first” so naturally we couldn’t do something that would increase their costs so substantially.
But patrons loved their creators and perhaps they’d be willing to shoulder the processing fees on the creator’s behalf. We knew that if we asked patrons they’d say “of course!” even if they didn’t mean it because asking people to predict their future decisions is full of problems. Instead, in our infinite wisdom, we thought the best way to find out for sure would be to run a controlled experiment and measure negative conversion impact.
We wouldn’t actually charge patrons a processing fee; we’d just show them that fee in the checkout and observe (across thousands of potential subscriptions) what would happen to their conversion rate. And because the fee dollar amount would increase or decrease depending on pledge size (2.9% + $0.35) we also wanted to study the impact on pledge size, conversion by pledge size, and a whole bunch of other dimensions.
So this is exactly what we did—structuring a controlled, A/B test to reach a statistically significant conclusion on pledge impact if we moved fees from creators to patrons. A classic example of quantitative discovery. We knew that conversion would decrease but expected the increased earnings from creators would outweigh that. And it turned out: we were right. There was some degradation in conversion but creators would end up making more money (albeit from fewer patrons).
Armed with this new knowledge we were ready to tell the world and pre-announce that these changes were coming. A/B testing for the win! Ship it!
On December 7th we pre-announced to creators, via email…
In my article “Beyond NPS - How to Build and Measure Trust with your Customers,” Vanessa Van Schyndel recalls how this landed:
“In one particular fumble, now remembered in infamy in creator circles across the internet as “The Fee-asco”, we announced that we were going to charge patrons an additional service fee per pledge. The rationale was that by charging patrons a bit more per pledge, Patreon could cover the costs of running the business while also giving creators a larger cut. Overnight, the announcement caused enormous backlash. Twitter was a dumpster fire. The support queue was full of messages ranging from confusion to violent threats. Creators were deleting their accounts left and right. Creators were furious at Patreon for stepping in between them and their patrons.”
First lesson learned: when you’re doing something sensitive like changing prices for existing customers (and those people have the largest reach on the internet and most engaged followers) you might want to actually run it by them first.
When the internet imploded with the anger and fury of over 100,000 creators we followed that up with an updated blog post further explaining our decision. Look, I found it!
By all objective measures it was a good post. It laid out the logical reasons why we were making the change, why it was good for creators, how they’d make more money, how we had tested it extensively “for over a year,” and why it was necessary for the continued evolution of the platform.
We had taken the approach popularized a song by The Animals:
“I'm just a soul whose intentions are good…
Oh Lord, please don't let me be misunderstood”
Six days later we announced that we were reversing our decision and not rolling out the fee change.
Ouch.
So what went wrong here and why did we have to roll the anticipated change back 5 days before it was set to go live?
There were six primary mistakes we made that can be bucketed into three themes:
Mistakes with the solution
Mistakes with the discovery
Mistakes with the communication
Mistakes With The Solution
The solution itself was flawed. Imagine walking into a store and as you’re approaching the checkout a tiny gremlin jumps in front of you and demands you pay him 2.9% + $0.35 before you can leave the store. The store owner has no idea where that gremlin has come from or why they’re demanding payment.
The store owner wouldn’t be very happy with that gremlin.
We were the gremlin. Unsurprisingly the 100,000 store owners on Patreon weren’t very happy with this solution and primarily for two reasons:
We were charging fees to their customers. Much like the gremlin above to the store owner, the creator viewed patrons as their customers. We were supposed to be the platform in the background and had violated that tenant of our “put creators first” policy while simultaneously thinking we were putting them first by reducing their fees. They didn’t care.
Capturing more than we created. Even though we thought we had done something great here we hadn’t introduced anything of value in this update except for some cost savings for creators. But while they were saving money, they were going to lose some patrons. It turned out that they cared a lot about that and a lot less about the cost savings or revenue increase.
And we would have heard both of these problems if we hadn’t made mistakes with the discovery process.
Mistakes With The Discovery
There were two primary mistakes we made here—the first was that we had a general lack of discovery outside of a (quite rigorous) series of A/B tests. And while those tests will tell you, quantitatively, just how good or bad something could be it won’t tell you much about the underlying reasons why. In this case, we didn’t bring this proposed solution to creators to ask them for their feedback. We just ran it as an experiment, understood the sensitivities, and thrust it upon them like the gremlin in the store. Only after we announced the pending change did we get a torrent of feedback that showed our obvious (in hindsight) blindspots.
The second mistake was that we allowed the quantitative to trump the qualitative. We didn’t have enough (or even any) conversations with creators to ask them about this change. We had plenty of conversations about the things that this change was supposed to unlock: anniversary billing, reduced fees, etc. but none about what changes would have to occur (we believed) to get us to those outcomes.
And once we realized we had dug ourselves a hole with our communication… We just kept digging.
Mistakes With The Communication
One of the things I learned in working with a lot of creators over the years is that they tend to respond to emotional appeals more than logical ones. Lest I be vilified for this comment, let me clarify: when making an argument there are three types of appeals—logical, emotional and ethical—and in my experience creators were able to best connect with emotional appeals rather than logical or ethical ones. This does not mean that creators are illogical (far from it) just that in this case we were trying to explain the details of the decision, the painstaking experimentation, the cost savings, etc. and they didn’t care because they were very concerned about the reaction that their patrons would have. We were making a logical appeal when they needed an emotional one.
But that wouldn’t have mattered much because we didn’t have any reference customers for this solution. Again, had we done discovery and come up with a solution that creators were excited about, we could have leveraged some lighthouse creators as reference customers—those willing to speak about the solution publicly and praise it. We didn’t have any. What’s worse, because we had relationships with lots of creators we were hearing privately that they understood why we were doing it (logical appeal) but they couldn’t publicly support it because it was like throwing their patrons under the bus. So we had no one willing to go to bat for us.
This pricing experience taught us some valuable lessons and in 2019, the next time we announced a price change, we nailed it.
Case Study 2: Win Customer Support and Build Trust – What We Did Right at Patreon
After the painful lessons at the end of 2017 we knew that we had to do something different and better the next time. We also knew that foisting changes on patrons was not a tenable solution. Our conversations with creators following the “fee-asco” caused us to change our thinking on the solutions to unpredictability, scale and even our approach to the membership model.
But there was another growing realization that we had in 2018; our ambition was growing faster than our bottom line. We had creators with varying levels of experience and sophistication, highly differentiated needs, and very different expectations around the types of features and level of support they wanted from the platform. And we were charging them all the same, low 5% platform fee which was anchored to the crowdfunding model of Kickstarter we had said goodbye to in mid-2017.
We decided to punt the membership model question—Patreon would eventually roll this out in September of 2022 (!!!), over four years later and instead address predictability, scalability and differentiated needs.
This time we approached the problem much differently by starting with both qualitative and quantitative pricing discovery. We conducted a broad-scale survey of current and potential creators, bucketed by different earnings types, number of patrons and audience size. We deconstructed our product into hundreds of individual features (added some new ones) and used a conjoint study to both identify optimal feature bundles as well as whether any features would reduce desirability of a bundle. That gave us some quantitative measures and helped us construct various pricing packages. [Note: I wish I had a copy of this study, but alas I could not find it].
Next we talked to close to 1,000 individual creators and showed them the various packages we had constructed to get their qualitative feedback.
Between these two studies we were able to identify some interesting patterns:
Creators who were not already on Patreon were relatively price insensitive. Most were used to the rest of the creator industry which charged fees upwards of 30-50%.
Those who were already on the platform were more price sensitive, but still wanted a whole slate of features that were unsustainable at a 5% fee structure.
Certain functionality was table stakes though creators at different stages clearly wanted different things. Generally speaking, those earlier in their career just wanted to get started; those at the midpoint wanted growth and those farther along wanted support managing a more complex business (and tools to translate complexity into revenue). They were all drawn to feature packages that included stage-appropriate functionality.
This was a multi-month exercise but provided a lot of clarity and confidence that we better understood what current and prospective creators wanted and would pay for. We settled on a three-tiered approach - 5% for a new “lite” plan, 8% for a “pro” plan and 12% for a “premium” plan. We also decided that we would package a new, fixed rate processing fee for creators and introduce a micro-transaction processing fee for any payments under $3. These fees would continue to be borne by the creator but they would eliminate the fluctuation in processing fees that creators saw every month.
You might wonder why we decided to make all these changes at once. First, through our research we realized that creators were less sensitive to the actual processing fee percentage and more sensitive to the lack of predictability. We had previously thought that eliminating processing fees altogether and passing them to patrons was the right approach and learned that it was decidedly not. Second, if you’re going to make a change to pricing it might as well be all at once and not have to make multiple announcements about pricing.
The final test was finding and winning the support of several reference customers.
We had already determined, in our conversations with existing creators, that there would be a legacy pricing tier (we called them “founding creators”) for those who didn’t want any enhanced functionality and had launched their patreon page prior to a particular date. For those who did want something more, we offered the 8% and 12% pricing tier they could upgrade to. So we were looking to win support for the 8% and 12% plans amongst our existing customers. We wanted to avoid the lack of public support we had experienced during the 2017 “fee-asco.”
To establish our reference customers we created a target list of dozens of the most connected and influential creators on the platform and conducted 1-on-1 conversations with them. We explained in simple terms what we were planning, that nothing needed to change for them if they didn’t want it to, and that we’d be releasing new functionality alongside the new pricing while stabilizing their fluctuating processing fees. The key part of the conversation was that we told them this would enable the long-term success of Patreon as a platform they could rely on. We were making an emotional appeal. Then we just shut up and listened. And the feedback was positive! There were some questions which we addressed and those questions helped us refine our messaging and positioning.
With months of qualitative and quantitative discovery behind us we were able to confidently proceed. We announced the new pricing in early 2019 and rolled it out in May of that year. And for once, the internet didn’t implode.
I’m going to skip over the ~10-12 months of development work to enable tiered pricing, lock and unlock features based on your pricing plan, build a stripped-down “lite” pricing tier, build team accounts, merch, special offers, a premium support and account management flow, and completely re-build the onboarding experience to accommodate price selection. That’s a topic for another newsletter.
Instead, I’m going to end with the four core lessons learned to soften the blow for your existing customers.
How To Land Price Changes With Your Existing Customers
Create more than you capture
One of the key differences between the 2017 and 2018/19 pricing stories is that we took our time and focused on providing more value to creators than we captured the second time around. We first studied the problem through various discovery techniques and then we actually built new, desirable functionality that didn’t exist in order to more-than-justify small price increases. We also worked hard to expose the dozens (hundreds) of features that already existed that often worked behind the scenes to power memberships.
If this particular lesson sounds familiar it’s because it also worked in the 2017 rebranding and repositioning of Patreon. In that case we justified our “pivot” to Membership by demonstrating it with new features and functionality.
Notify, notify, notify
We talked to thousands of creators about the second price change. We built reference customers and won support for the changes, listened to feedback, and incorporated it into our plans. When we were ready to roll it out we gave creators months of notice and allowed them to launch ahead of the price increase to participate in the “legacy” price.
Perhaps most importantly for the existing creators we were consistent in our communication that nothing will change for you unless you want it to. We repeated that message an untold number of times to every existing creator we talked to. And we saw creators share that same message to each other when the announcement happened.
Legacy / founding creator pricing
I cannot stress enough the importance of providing your existing customers with some sort of grace period or permanently frozen pricing package for placing a bet on your company at an earlier stage. In most cases I’d recommend ~6-12 months. Patreon actually did it forever. There are still some creators leveraging the “founding creator” pricing package and paying 5% for the equivalent of the “pro” tier. Five years later.
Of course, your mileage (and unit economics) may vary. But this is one of the reasons that creator trust of Patreon remains very high.
Provide a “downgrade” alternative
When we launched the Pro and Premium pricing plans we also introduced a “lite” tier. In October ‘23 Patreon deprecated that tier (4+ years later) but at the time it was a way for new creators to enjoy the same “founding creator” pricing if they didn’t want all of the functionality offered in the higher price plans. Even creators who switched to the higher priced plans and lost their “founding” status could revert to the lite plan.
Netflix has taken a similar approach by offering the ad-supported plan. It’s a downgraded option for sure (I hate ads) but for those who don’t want to participate in an increasing price you at least do have another option. The same is true with Shopify who offers a $5/month starter plan which is a pretty steep price downgrade from their basic plan. It may not work for you, but it does exist.
Wrapping It All Up
So now you see a tale of two price changes—one that went about as bad as it could have and one that landed perfectly.
In the first change we made several large mistakes around the solution, the discovery, and the communication. In the second, we learned from the past rebalanced the value creation and capture equation, worked with creators from the beginning and over communicated, provided legacy pricing and offered downgrade alternatives.
My only criticism of our approach the second time was the pace. It’s understandable why we moved so methodically: we were much more cautious after the “fee-asco” and wanted to ensure that we were buttoned up on this one-way door decision. We also had a lot to rebuild to get the pricing to work properly. Still, I think we could have made some decisions faster.
In the end it worked out well for us. Patreon doubled its margins and was able to invest more to serve the needs of creators as a result.
If you have a great (or not-so-great) pricing story I’d love to hear about it in the comments.