Min-Kyu Jung Subscribe

Sign up for new posts in your inbox

If you just want to say hi, my email is mkyujung@gmail.com

How I would grow Gumroad

“It doesn’t matter how amazing your product is, or how fast you ship features. The market you’re in will determine most of your growth.” - Sahil Lavingia, Gumroad CEO

Let’s say Sahil is right. If Gumroad’s growth is restricted by the size of the indie creator market, then one way for Gumroad to ensure its future growth is to actively grow the indie creator market itself.

Stripe has the ambitious goal of “increasing the GDP of the internet”. They claim that this is because “removing the barriers to online commerce helps more new businesses get started, expedites growth for existing companies, and increases economic output and trade globally”.

This means that Stripe does things that don’t meaningfully grow Stripe directly. Stripe Atlas, for example, helps founders navigate the red tape involved in forming a new company, and Indie Hackers helps independent creators build businesses with the help of a supportive community.

These initiatives don’t provide immediate, quantifiable results in the way that a paid acquisition campaign would. In the long run, though, this strategy expands Stripe’s total addressable market, and beneficiaries of this market growth are disproportionately likely to use Stripe rather than another payment solution.

Stripe’s mission is audacious. Sure, Stripe is a $35b+ company with truckloads of capital for moonshot projects, and Gumroad isn’t. Even so, I think that Gumroad could frame its growth strategy in a similar way.

How can Gumroad increase the GDP of the indie creator economy?

This post is focused mainly on the distribution side of growth. See How I would grow Substack or The future of contracts is code for my product-centric posts.

As part of my research for this post, I reached out to four Gumroad creators who answered my questions on their experience as a creator. I’ve included snippets of their responses throughout this post.


I. Gumroad’s #1 metric

Gumroad helps creators do more of what they love.

One way to quantify how effective Gumroad is at helping “creators do more of what they love” is volume processed. So this will be the top-line number that I focus on in this post.

We can break volume processed down into two buckets:

Volume processed = volume processed per creator * number of creators

We can then break each of these numbers into two further buckets:

Volume processed per creator = price per unit * number of units sold per creator

Number of creators = acquisition * retention

Breaking down our volume processed metric like this helps us understand the various levers that drive that number. I’m going to assume that Gumroad can’t do much to effect the price per unit sold, because that’s set by the creators. That leaves us with three levers that are within Gumroad’s control:

  • “number of creators”, which can be further broken down into acquisition and retention; and
  • “number of units sold per creator”.

We can visualize this as a “volume processed” framework, with the blue circles representing levers that can be moved:

In this post, I’ll go through each of these levers in turn. But first, let’s talk about how creators create content.


II. The four phases of a content creator

Making content is a long process. Gumroad makes distribution easy, but by the time a creator gets to this point, there’s already been a long timeline of events that have occurred. Usually, the only visible part of this timeline is when the product is released and distributed. I like the analogy of an iceberg to describe this phenomenon.

I’ll use myself as an example. Six years ago, I liked making music in my spare time, so I decided to make an EP. Here’s what that process looked like:

Phase 0 – Default state (99.9% of my life)

  • No or little interest in making music

Phase 1 – Base-building (1 year)

  • Learning about music production
  • Finding production equipment
  • Learning production software

Phase 2 – Content creation (4 weeks)

  • Composing/songwriting/arranging
  • Recording music
  • Mixing – EQ levels, compression, etc.
  • Mastering

Phase 3 - Release and distribution (1 day)

  • Putting music on the internet
  • Marketing

In my case, I was just a kid messing around with a guitar and second-hand equipment in my bedroom. “Releasing” my project meant uploading it on Bandcamp and “distributing” it meant asking friends to check out my mixtape. I made $5.00.

But even if I had grander ambitions, releasing the project would still have been by far the least difficult and frictionless aspect of the process. Gumroad and its creators only get paid in Phase 3, but most of the obstacles in the creative process are actually in Phases 1 and 2.

The further down the iceberg you go, the harder it is to effect change. Phase 1 is harder to effect than Phase 2, and Phase 2 is harder to effect than Phase 3. And getting people from 0 to 1 at scale is nearly impossible.

So it makes sense that Gumroad would start by focusing at the top of the iceberg, rather than the bottom. Phase 3 strategies are focused around helping creators distribute their products — say, for example, Gumroad’s “Discover” feature and reorganizing the onboarding flow so that it’s easier to upload content. However, focusing solely at the top of the iceberg limits growth because most of Gumroad’s would-be creators are further down.

With this in mind, let’s think about how Gumroad could acquire and retain creators stuck in Phase 2 (and earlier phases).


III. Lever 1: Acquisition

Gumroad doesn’t spend any money on customer acquisition. Instead, Gumroad’s customer acquisition process looks like a loop:

  1. New creator joins
  2. Creator sells content
  3. Other creator see that a creator is selling content on Gumroad
  4. New creator joins

One important aspect of this growth loop is that it doesn’t just acquire creators who are making decisions right now about how to distribute their products, but it acquires future creators who are still stuck in Phases 1 and 2. If a future creator is browsing their Twitter feed and finds someone selling, say, photo sets on Gumroad, they’re more likely to use Gumroad to create their own photo sets when they eventually move to Phase 3.

“Most professional artists I follow typically have mentioned Gumroad at one point or another when selling their products … so I just went with the most popular option whose name I heard being passed around the most.” Meg Syverud, creator of “A Study in Scarlet & Blue: A Spider-Man AU Comic”

This loop is a great growth engine, but Gumroad can find more opportunities for growth by framing acquisition in different terms: how can Gumroad move creators from Phase 2 to Phase 3?


Gumroad generated content

I’d like to see Gumroad experiment with building its own content aimed at helping creators move from Phase 2 to Phase 3.

I’m sure that content marketing isn’t a new concept for Gumroad. For starters, like most companies, they have an active blog. I’d also consider Sahil’s Twitter to be a form of content marketing — most of the Gumroad creators I’ve spoken to have mentioned Sahil’s tweets as playing a part in them discovering Gumroad.

“I don’t know if I can pinpoint [how I found out about Gumroad]. I’d definitely seen Sahil in the community via Twitter, blog posts, and podcasts.” Alex DeBrie, author of “The DynamoDB Book”

But I’m not sure if Gumroad has tried to create content with the express goal of moving creators across phases. There are two requirements for this kind of content. It needs to be (1) useful enough to meaningfully impact on creators’ movement between phases and (2) interesting enough to actually be consumed.

From Priceonomics’ content marketing handbook — replace “information that helps your company” with “information that helps creators create content” and you have the intersection that Gumroad should be aiming for. Incidentally, Priceonomics’ content marketing handbook is itself a great example of the type of content that Gumroad could produce.

When I was in Phase 2, I sought out videos on music production and blog posts on affordable home studio setups. If I made my EP with the guidance of a full-length Gumroad handbook or video course covering the entire “Phase 2” hobby-album production process from beginning to end, I would have been all but guaranteed to have used Gumroad as my distribution platform. Also, importantly, it could have been the difference between releasing my project and giving up before getting to Phase 3.

One problem with this idea is that it’s hard to test quantitatively. It’s really difficult to definitively establish that a new creator joined Gumroad because of a piece of content that they consumed.

So here’s how I would experiment with content. I would deep-dive into three of Gumroad’s biggest or most promising categories and talk to creators to understand the process of moving from Phase 2 and Phase 3 within these segments. Then, I would produce a comprehensive piece of content for each of these categories. I would take advantage of Gumroad’s internal data and access to creators to produce interesting content that nobody else can.

Here’s a few examples of the type of content Gumroad could create:

  • How to Create, Design, and Release your Own Ebook
  • How to Create a Video Course from Start to Finish
  • The Complete Guide to Creating a Music Album

These examples are pretty generic. The more longtail Gumroad goes, the more likely it is that they’ll create valuable content that isn’t easy to find elsewhere. Too far longtail, though, will mean that the addressable market for their content will be too small to be worth the investment.

A few months after this content is created, I would test a new cohort of creators to figure out (1) how they heard about Gumroad, and (2) why they became a Gumroad creator. I could then measure what percentage of those creators joined Gumroad because of its content, which when extrapolated, would provide an approximation of the total number of new creators being acquired through Gumroad content, being “X”. If the investment in content was worthwhile based on X, the experiment will have been successful.

The hypothesis would look like this, with square brackets denoting variables to be decided:

Hypothesis: Within [three] months of releasing three guides, these guides will have resulted in the acquisition of at least [X] new Gumroad creators.

This still isn’t an ideal test because the feedback loops are so wide. It would take a long time to attribute any growth to Gumroad’s content, and it’s complicated by the fact that evergreen content will continue to acquire creators long after it is published.

But while testing this hypothesis isn’t easy, Gumroad should still be able to receive plenty of feedback on its content pretty much immediately. If people are saying things like “wow great guide!” that suggests that the guide is good. And if a lot of people are reading a Gumroad guide, that at least tells us that the guide is being seen by many potential creators at the top of the funnel.


IV. Lever 2: Retention

How do you define retention for Gumroad? One way to do this would be to define a creator as being “retained” if they are subscribed to Gumroad’s premium service, and churned if they cancel their subscription.

I don’t think this is the best way to frame retention because Gumroad isn’t a pure SaaS business. Only a portion of its revenue is derived from subscriptions; and remember, we’re optimizing for volume processed, not subscription revenue.

Instead, Gumroad’s retention can be measured in terms of how often a creator comes back to publish a new product. Gumroad is different from a typical SaaS business where revenue per user is constant or increasing over time. Even with Gumroad’s premium subscription offering, their revenue per user comes up and down in cycles centered around product launches.

This means that Gumroad needs to constantly be moving creators from Phase 2 to Phase 3, even after they’ve been acquired.

I’ve described the process of moving through phases as an iceberg that looks like an inverse sales funnel. However, it would be more accurate to describe it as an ongoing, cyclical process. Once a creator creates a new product, they’re not “done” making products. They’re not stuck in Phase 3. They go back to Phase 2 or 1 and start again. Experienced creators have tighter creation cycles because they don’t need to go through Phases 0 and 1 to same the extent that a new creator would. But they still need to keep creating.

This is why I like the idea of framing Gumroad’s growth strategy as “move creators from Phase 2 to Phase 3” — everything I covered when talking about acquiring new Gumroad creators also applies to retention.

Product changes to drive retention

The best way to move retention isn’t to focus on retention itself, but the levers that drive retention. The main lever that drives retention is activation, which is a growth marketing buzzword for when a user discovers and accesses the core value proposition of the product.

This means that we should try to understand how Gumroad activates its creators. One way to do this without quantitative data is to talk to creators directly. When I did this, I asked them why they used Gumroad, which was a plain way of asking what Gumroad’s value proposition was for them.

The answer is that (1) Gumroad helps creators make money (2) Gumroad is really easy to use and (3) Gumroad lets creators keep control over their content. So Gumroad’s goal is to figure out how to deliver these value propositions to creators as quickly and as frequently as possible.

“Gumroad is user friendly, and they charge fair prices. Teachable for example, which I’ve also used, charges more. Gumroad is a good way to have a more or less independent platform, as opposed to say Amazon.” P.D. Mangan, creator of “20-Minute Home Fitness: The No Gym Alternative”

With this in mind, what is the activation point for a Gumroad creator? A Gumroad creator can’t experience all three of Gumroad’s value propositions until at minimum she makes her first dollar on the platform. It’s the earliest moment where she can say “Wow! This Gumroad thing is really great!”

Because of this, I think that Gumroad will find that if new Gumroad sign-ups make their first dollar within a certain timeframe, they will be significantly more likely to become retained creators.

Hypothesis: Gumroad sign-ups who make their first dollar within [X] days are [Y]% more likely to publish another product with Gumroad.

Once this hypothesis is tested and confirmed, Gumroad can form a north star metric to drive its product decisions: # of creators who earn their first dollar within [X] days of sign-up. If Gumroad can confirm its first dollar hypothesis, this metric would be a leading indicator of volume processed, and therefore an effective rallying point to frame product development around.


V. Lever 3: Number of units sold per creator

The last lever left in our volume processed framework is the number of units sold per creator. This lever is mainly a Phase 3 problem.

Gumroad is already moving in the right direction here. I expect that their upcoming Memberships feature, which will let creators charge a recurring subscription fee for their content, will increase revenue per user.

Gumroad’s Discover and Posts features also help with discovery and converting leads, which increases the number of units sold per creator. Let’s talk about these two features.

Discover

Gumroad’s Discover feature is a page that helps audiences find Gumroad-recommended products. In exchange for these recommendations, Gumroad takes an additional 10% for each sale made through Discover.

“The only wish that may be here is that Gumroad [itself] also promotes [my] products, and helps with sales. But in fact, as far as I can see from the analytics, they already do it. Of course, not so many, but there are several sales that Gumroad arranged [themselves], and took a small percentage for it.” Daniel Mikhailov, creator of “The Key to understanding Hand Drawn Animation”

I don’t know how Gumroad chooses which products to feature on its Discover page. Whatever algorithm they use, it seems to be heavily weighted towards popular content. This makes sense; a product that thousands of people have bought is usually more useful to the average customer than a product that no one has ever used before.

One thing that Gumroad could experiment with is using its Discover feature to help activate new creators. If Gumroad can establish that its north star metric should be “# of creators who earn their first dollar within [X] days of sign-up”, then it might pay to tweak their algorithm so that it features creators who (1) have published their first product or project on Gumroad within X days and (2) have yet to earn their first dollar on Gumroad. These creators are particularly vulnerable to not moving back to Phase 3 if their first launch fails to get a single sale, and Discover might be Gumroad’s best chance of activating these creators before they churn.

This needs to be balanced against the risk that the average quality of products on Discover suffers as too much unproven content finds its way onto the page. This balancing act is hard, because the distribution of successful Gumroad products almost certainly follows a power law, where the majority of creators will have barely any sales.

Ideally, Gumroad would be able to find leading indicators for products that meet the above criteria but are also likely to be high quality and successful, if only they were given a small push.

Posts

Gumroad lets creators make posts on their personal Gumroad pages that get sent straight to subscribers’ inboxes. I’m optimistic about Posts. It’s a step towards turning to Gumroad into an all-in-one landing page for creators.

According to this screenshot from Gumroad’s January 2020 online board meeting, Posts had impressive early adoption from creators not long after launch.

However, it’s now coming up to a year since Posts has been released. I’ve looked at 30 Gumroad profiles (which is, admittedly, not a particularly sophisticated way of getting results, but hopefully directionally correct) and have come to two conclusions:

  1. the vast majority of creators (26/30) don’t have any posts on their page, and of those that do, most (3/4) have only one or two posts; and
  2. posts are mainly used by creators to advertise discounts on their products.

In other words, creators aren’t using Gumroad’s Posts feature like they would use Medium for publishing blog posts, or even like they would use Substack to publish newsletter content. Instead, Posts are either not used, or they are used like advertising collateral to convert leads.

One reason I’m not super excited about this use-case of Posts is that I don’t think it’s a huge value add for subscribers on the receiving end of these emails. I get that if I’m planning on buying a product from a creator, then it literally is a “value add” for me to get a discount code for that product in my inbox. I also understand that if I have voluntarily entered my email into a creator’s Gumroad landing page, there is a non-trivial chance that the previous scenario will apply to me, especially with Gumroad allowing creators fairly granular control over what kind of people their posts are sent to.

Still, this is outweighed by the fact that outside of this singular context, an email with a discount code is completely useless and even a little spammy.

Instead, the ideal use-case for Posts would be if creators used them to publish interesting content (that isn’t locked behind a paywall) to act as a distribution channel for Gumroad to acquire more creators and customers.

Earlier, I talked about how Gumroad should create content aimed at helping to get creators from Phase 2 to Phase 3. Posts mean that Gumroad can just leverage its enormous creator base to do this for them. Posts wouldn’t even need to be created for the express purpose of helping other would-be creators. The mere fact that useful Posts are more likely to be shared and be SEO friendly means that they would be a step towards growing the GDP of the indie creator economy.

Gumroad creators don’t seem to be using Posts this way. Trying to change the way your users use your product is very difficult, and probably bad practice. But I think that Gumroad should at least try, if they haven’t already, to understand why creators use the Posts feature the way they do; if they aren’t writing more substantive and useful content due to obstacles within Gumroad’s control, then it would be worthwhile trying to fix this by doubling down on Posts. It could become Gumroad’s most important feature.


VI. Growing the GDP of the indie creator economy

Gumroad is in a great position. It’s profitable and doesn’t have the pressure of hyper-growth that comes with VC funding. This doesn’t mean that growth isn’t important. It just means that Gumroad doesn’t need to look for growth of the “pay a dollar to earn 90 cents” variety. They can afford to grow sustainably by building equity in the indie creator economy over the long term.

Helping creators “do more of what they love” might not just be an idealistic tagline, but an effective growth strategy. Gumroad can do more than just acquire existing creators. Gumroad can create new ones.


Thank you to Meg Syverud, Alex DeBrie, P.D. Mangan, and Daniel Mikhailov for helping me with this post. You can check out their content here:

Thanks also to Jun Kim for feedback on this post.