Author Archives: Casey Winters

The Monetization Playbook We Used at Eventbrite

One of the problems we faced when I joined Eventbrite was that Eventbrite had a pretty low take rate, meaning when event creators sold paid tickets on our platform, we took a very low percentage. And when creators sold free tickets on the platform, which are the majority of tickets, we didn’t make any money. You need an incredible amount of volume to make profits with low take rates, like say Paypal or Stripe, and even though Eventbrite has tremendous scale, it wasn’t enough to make material profits. We needed a way to make more money for the value we created if we wanted to be a successful company i.e. a new monetization playbook. I’ll discuss below the playbook we built, and some of the process to get there.

A tweet I made when I was working through this at Eventbrite.

At the most abstract, there are two ways to grow profitable revenue for a network business:

  • Make more revenue per customer
  • Increase network volume

Eventbrite had tried to improve its revenue per customer before I arrived as Chief Product Officer. The company did what most SaaS companies do in this position: hire a consulting firm, work closely with them for a period of time, and decide which of their recommendations to implement. This created the first packaging model for Eventbrite. Instead of one package, there were now three that had different features for paid event creators. This all sounds very non-controversial, but the impact was pretty negative for Eventbrite. While revenue initially went up, over time, growth slowed as event creators eventually churned due to the new prices, which doesn’t just impact retention, but also acquisition, as the marketing they do for their events is the main driver of new creators. Revenue per customer went up but network volume slowed. Not a great tradeoff. In addition, no one chose the first plan, and free creators still paid Eventbrite nothing.

Eventbrite was pretty constrained on how to adjust levers in the pricing model. As a transactional business with payment processing costs, you tend to have both fixed and variable fees. As you drop one or the other, you basically decide whether you’re hurting pricing for event creators with smaller or higher priced tickets. Companies like Orb have since come along to make bigger changes in pricing more manageable, but our system at the time was very brittle and rigid.

So the company tried to re-ignite the network volume growth the pricing change hampered, but with a pricing model where two thirds of creators still don’t pay and a sub-10% charge for paid tickets that made many forms of paid growth unviable. 

The product team rushed in to create new value for event creators, but this just put more pressure on the take rate as you’re now trying to maintain more software for the same amount of revenue. Everything the product team recommended default shipped to all pricing plans as well, which meant we weren’t improving the value of more expensive plans over time. The baseline reasoning for this is to maintain pace with competition, which, if you hear it in your company, might mean we’re not investing in the right features that can’t create differentiation. But also I think this default practice reflected a lack of comfort with pricing recommendations and becoming more business oriented in our approach to product development. Or maybe just laziness.

This laziness extended beyond pricing the new features, but also marketing them. Many features lacked awareness with our creator base, creating a deadly loop of product development.

We needed to shift into a new model that forced product teams to think about both pricing and marketing features as a core part of their job, and it was of course my job to teach them how to do it. The model needed to look more like the below:

Obviously, you cannot charge for every feature. Sometimes, we do need to keep up with competition, but changing the default is important to getting a team to help the company actually become a sustainable business. 

If we get in the habit of charging for the value we create, the main question is how to do so at both a feature / product level and an overall offering level. The first question to actually ask is does the feature or product create value that is more important than how much we could monetize it. How we thought about answering this question is:

  1. Does it drive virality? If so, we’re incentivized to give it away.
  2. Does it drive activation to long term retention? If so, we’re incentivized to give it away until activation is achieved.
  3. Does it drive lifetime value? If so, compare lifetime value gains vs. willingness to pay to determine if we should charge for it.
  4. If something does not drive virality, activation, or lifetime value, but people have willingness to pay, we should charge for it.
  5. What is the cost to support this (eng costs, unit economics)? Does the value provided in the above questions cover for that cost? If not, don’t build it or sunset it. 
  6. Do the costs have economies of scale? If so, can sharing these economies improve lifetime value? If so, refer back to rule no. 3.

If we decide to charge for something based on answering the above questions, the next question is how. We used the following 2×2:

Low # Creators Use High # Creators Use
High Willingness to Pay Add-On Higher Package
Low Willingness to Pay Don’t Build Core Package

What creates low willingness to pay for things that a lot of creators will use is that they are available elsewhere in the market. While you of course need to build some of those, that is not where enterprise value is created. Enterprise value is created in the top row where there is differentiation and therefore higher willingness to pay. Answering these questions while working on the initial concept of a product vs. at the end when it’s already been built is a good prioritization element in and of itself.

The reason the 2×2 defaults to packages is that if you apply a willingness to pay framework too literally, monetization can become too complex. Pricing ever feature a la carte quickly turns into something approaching the toothpaste aisle at a drug store. In theory, there is a reason for why there are a hundred plus varieties with different features and price points, but no customer will go through the work to understand all of that. So there is still a bundling art form that product collaborates with marketing on to make the packages viable and digestible at perhaps the sacrifice of some individual feature willingness to pay. Frameworks are great for 80%, but I always preach that product folks can’t let the framework absolve themselves of responsibility for using their brain to make the company successful. 

If you want to learn more about pricing and packaging frameworks and how to measure willingness to pay, I highly recommend Reforge’s Monetization & Pricing program. If you’re not sick of me, the Pricing Strategy and Advanced Growth Strategy programs I helped create are also starting again soon.

Be sure to subscribe to my Substack to catch future posts.

Currently listening to my Dune: Part Two playlist.

Wood Grain Steering Wheels

I work with a lot of startups these days, and startups are frequently excited to talk to me about advanced growth ideas. This makes sense. I’ve been able to have success in my career leveraging advanced strategies to help startups win and to take more control over their growth. I built a course called Advanced Growth Strategy after all. In 99 out of 100 of these cases though, this is the wrong topic. Inevitably, when I dig in, any startup asking me about doing something advanced on SEO, activation, or paid acquisition is not doing something much more basic well yet, and the impact of the more advanced concept won’t bear fruit until they get some of these core concepts of growth working better.

I’ve been using this song lyric from Outkast more recently to snap them back to reality. “If you aint got no rims, don’t get no wood grain steering wheel, for real.” There is a hierarchy of concepts you should be focused on in building your startup, your product strategy, or your growth strategy the same way there is for pimping your ride.

Big Boi always had his priorities straight.

There are two reasons people start to do things out of order inside a company. The first is that many people just want to skip the boring building blocks to do something innovative, which is entirely backwards. Nailing all the fundamentals around building a product or a growth loop is what most of the time enables good innovation. If you don’t have strong product/market fit for your core product, an AI feature isn’t going to save that. Talk to customers! Why aren’t they sticking around? If your pages aren’t getting indexed and you have a bunch of duplicate content, don’t think about microsites or advanced link building ideas. Fix your site!

The second reason this happens is if you have worked on the basics in the past, you don’t have a good process to regularly audit whether they are still working as designed. You might abstract away the detail to a metrics dashboard and not actually run through the user experience regularly yourself, especially in B2B when you are not the customer yourself. Great product development teams get good at auditing what has been built regularly over time. You are surprised what you will catch, from entirely broken flows to slow pages to outdated designs. While fixing these might not 10x your growth, not fixing them will make 10x growth impossible over time.

Be sure to subscribe to my Substack to catch future posts.

Currently listening my Hip-Hop playlist.

Is My Second Product Failing?

In my recent essay about second products, I talked about how the goal of building a second product is not the traditional concept of product/market fit we’ve all been raised on regarding startups. In some ways, our goal with a second product is easier, and in some ways, much harder. The goal is to inflect the growth of the entire company over time. This is hard, and does not usually unlock easily. This can make it tricky to evaluate as an engineer, designer, product manager or CEO if the investment will bear fruit and whether one should keep going. I’ll explain how I think about these decisions as a product leader and give more context on the journey of developing entirely new products inside a company.

First off, let’s spend a moment on this product/market fit point. If the job is to inflect growth for the overall company at some point in the not too distant future, product/market fit in the general sense may not be enough. Let’s say your second product builds a new product that is growing healthily with a smaller group of customers than the core business at a much lower price point. By general standards, the team found product/market fit, but will never be able to inflect growth of the combined company. This is failure inside a medium to large company. 

When the above example is interpreted correctly inside a company, this means teams start to take larger swings that, if they work, can possibly move the entire business eventually and open up new growth opportunities for the entire company. Bigger swings generally mean a longer amount of time building toward something that could work in the future. How much time and money has Meta spent on VR? Amazon on Alexa? Google on GCP? Extreme examples sure, but partially why they are extreme is these companies’ new product development has to move revenue growth at what are now some of the largest companies in the world.

While we are used to startups needing a decent amount of time to find product/market fit (sometimes taking many years), companies that are scaling do not generally think on such long timelines. OKR’s, which most teams use inside scaling companies, are generally measured in cycles of three to six months, which is much too short for a team investing in new products to evaluate success. 

A good team will push the company to give them a longer time to journey toward a new product that can be a success. But this just creates another problem. For a team working on what is usually considered one of the most important projects at the company, executives want to hear about “progress” or “milestones.” Are we on the right track, or are we lost in the wilderness?

Good teams working on new products will outline the assumptions they have to prove, in order, and a rough timeline on when they expect to either prove or disprove those assumptions. It’s a shot in the dark, but for territory that is yet unmapped, it gives the rest of the company some waypoints as well as a general estimate of how long the team expects to reach the destination. Even though the team usually knows nothing in reality.

I liken this journey to the myth of finding a rainbow that leads to a pot of gold. We’re saying the rainbow starts here, we expect it to take us through these peaks and valleys, and we think the journey will take a while with lots of unknowns. The most important thing for a team to do at this stage is state what the destination should hold. 

The company should know at this point it’s a potential new product with a large enough addressable market, price point, and whatever other strategic dynamics are important to the company that can lead to a new S-curve of growth. Much of the time, this list includes certain weaknesses of the current product suite the new product hopes to mitigate. Such examples are expanding the addressable market for the company, increasing frequency of usage, improving unit economics, etc. This list is important, but often missed in the new product development journey because product teams want to see where their insights take them. I would argue this is a mistake. The destination has to be worth it for a team that may spend over a year developing something new, so we need to align on what “worth it” looks like.

Okay, so you’ve mapped a destination representing a huge pot of gold awaiting the company when you reach it. You also need to map the milestones, usually proving/disproving assumptions in a certain order, and what you expect to achieve along the way. Updates to executives typically focus around these points in time. It’s a convenient way to discuss everything you’ve learned that may be important while getting a general pulse check with the executive team on their excitement around what the team has been working on.

Image by MariaAllen via Midjourney

Inevitably, some of these assumptions will be wrong. Uber thought the best way to launch a food delivery product was to copy some earlier stage startups called Spoonrocket and Sprig that had a limited menu of food that could be delivered in ten minutes. It turned out copying Doordash and Postmates was the more viable option. This sort of pivot is common, and still got them to the destination of a growth market with a large TAM that offset peak times for driver demand and allowed new types of supply that weren’t a great fit to core Uber to also make money driving. It also increased lifetime value and unit economics for the combined business. Pretty nice win.

When these assumptions are wrong, it does not mean the team has done anything wrong. It means the team is developing a feedback loop with the customer and the market and a learning culture inside the team. The waymarkers were wrong in the case of Uber Eats, but the destination was still correct, and the team eventually found a more accurate map to the destination. This is generally how it’s supposed to go.

But what if the learning culture you’ve developed inside the team and the feedback from the market and customer base is not getting you closer to the original destination? This is where things get concerning, and teams get confused on how to move forward. There are a few ways in which this can happen:

#1 The destination is further away than we originally imagined.

In this example, there are many more waypoints the team needs to travel to to reach the destination. Think of it as the road being much longer and windier to get to the pot of gold. At this point, teams need to remap the path, and examine whether the larger investment of time and resources is worth it to reach the destination. Much of the time, the project will get too expensive and time consuming for too small a reward. This happened to us at Pinterest. A team inside the company started building a Q&A product around the Pins people saved to the network. With every investment the team made, the path to success lengthened in many ways, such as the amount of moderation and quality control required when Pinners on the network started asking more questions and expecting serious answers. The team eventually recommended sunsetting the project, and they did. Other times, the reward is still deemed to be worth it, and the team keeps moving.

#2 The destination is the wrong destination or does not exist. 

This is a common problem when the new product investment is too business focused and not enough customer focused. Bigger companies can get attracted to product ideas that the band Soundgarden coined “pretty nooses”. As lead singer Chris Cornell (RIP) explained, a pretty noose is “just sort of an attractively packaged bad idea, pretty much, something that seems great at first and then comes back to bite you.” I’m sure you’ve seen them in every Y Combinator batch and in every product brainstorm. Everything the team learns as they work toward proving their assumptions on these ideas is that the market just isn’t there or that the customer pull is not strong enough. This happened with Tinder Social. Matching groups of friends just wasn’t a problem people were that interested in using software to solve. Acquisition is very hard, retention is even harder, and the market seems incredibly small and difficult compared to growing the core business.

#3 The team has veered off track to a new destination.

This problem occurs when the team investing in building new products is too focused on the lean startup path of developing products and fails to remember they are not their own startup, but a team attempting to accomplish a business objective for a larger company. In this scenario, the company has to ask how attractive the new destination that seems to have more traction than the original destination is. I was working with a company recently working on a new product that would address a new market that was higher frequency. All of sudden, the team working on it veered back toward the market for the original target market, for which this product would have too small a TAM and not address any of the new market or frequency objectives we originally had for starting the project in the first place. We decided to course-correct.

#4 We cannot seem to get closer to the destination over time.

This can much of the time be one of the other failure modes, but it’s important to call out on its own. If the only thing the team learns over time is “well, that didn’t work either”, it may be time to shut down the project. New activity on the problem needs to be generating new learning that makes the map to the destination clearer. If all we are learning is we are still going in the wrong direction over time, the chances of finding the right direction trend to zero.

Development of new products with new value props is difficult, and the frameworks for building startups or building other products inside the company often fail to apply to these situations. This makes it super important that teams map their destination, map the territory they expect to go through and map the process they will use, and stay on track for all of those to ultimately make the best decisions to maximize an outcome that meaningfully impacts the company.

Be sure to subscribe to my Substack to catch future posts.

Currently listening to my Footwork playlist.

The Best Way to Drive Demand in Marketplaces is Hiding in Plain Sight

This essay was co-created with Dan Hockenmaier.

In marketplace businesses, the network is the product. If you’re not growing both supply and demand, the product generally isn’t getting better over time. And your network effects which compound organic growth are not getting better, and probably getting worse.

Most marketplace founders and leaders intuit this over time, so they obsess over growth metrics. In this essay we explore one channel that helps many marketplaces scale faster than others: supply driving demand. 

If a marketplace has the potential to use this channel and doesn’t, it leaves them unoptimized and susceptible to competition. But if they use it too much, it is no longer a classic marketplace and loses the compounding benefits of network effects. There is a Goldilocks Zone:

Driving between 10% and 40% of demand from supply is the Goldilocks zone to maximizing value of our a marketplace’s supply to grow demand. Obviously, other factors besides this determine the full value of a marketplace.

Exploring each case

Supply driving too much of demand (75-100%)

As Casey and Gilad Horev discussed in a previous essay, there are different types of marketplace models. In the SaaS-like network model, the supply is in charge of generating the majority of the demand, and the product acts as mainly a fulfillment or monetization vehicle for the supply. These products generally do not generate cross-side network effects. Because demand only comes to the product when supply tells them to, the product doesn’t feel materially better over time. Acquiring supply doesn’t get easier over time, because the product isn’t generating demand that supply has to come to the platform to access.

These models typically range from 75%-100% of their demand coming from supply, and include companies like Substack, Square, Eventbrite, Mindbody, and many others. Sure, they may generate some of their own demand over time, but not enough to unlock the types of cross-side network effects that power the best marketplaces to massive scale like Airbnb and Doordash. Generally, part of their strategy to grow is actually driving more demand that does NOT come from supply over time, and their value tends to come much more from how well the SaaS business can upsell new products over time.

The 50% threshold

Our rule of thumb is that a marketplace needs to generate at least 50% of the transactions for cross-side network effects to exist. If a supplier would get 50% more transactions through a platform vs. on their own, most rational suppliers would prefer that vs. the 15-25% higher margins they could get going directly to their customers and avoiding a marketplace fee.

This creates that attraction of supply to the product, decreasing acquisition costs. But it also means the selection for demand improves more quickly, making for better discovery, higher conversion, and generally higher frequency on the demand side. We see these elements in the best marketplace businesses. They have amazing cross-side network effects.

Supply driving very little demand (0-10%)

This isn’t ideal for the simple reason that supply driving demand is a great channel if you can get it to work. It is generally highly scalable (meaning its potential as a channel increases as supply grows) and relatively cheap (you are paying to incentivize your suppliers, which is generally cheaper than buying ads on Google or Facebook).

Can you build a great marketplace without supply driving demand? Of course. The company will just need other acquisition strategies they can leverage to be successful, such as virality, paid acquisition, sales, and user generated content, and they might need to spend more money to get to scale vs. marketplaces that can get demand side growth from supply.

The Goldilocks Zone (10-40%)

So, what is ideal? “Just right” appears to be driving your supply-led acquisition channel to north of 10% of demand, but less than about 40% where you start to see too much weakness in the company’s own demand drivers and resulting impact on its network effect.

(As an aside, the percentages in this essay should be thought of as transactions, not customers. If the suppliers are driving >50% of buyers in a marketplaces, but the marketplace is able to effectively cross-sell them to other suppliers on the marketplace such that transactions from this channel are <40%, that is likely just fine.)

At Grubhub, restaurants directly or indirectly drove about 30% of new demand-side acquisition at scale (and Doordash and Uber Eats have replicated this strategy to likely similar percentages). At Grubhub, we gave the restaurants a lower take rate for the orders they drove, but even so, restaurants driving customers to Grubhub became one of our most cost effective acquisition channels. At Faire, both sides of the marketplace refer their existing customers onto the platform because it is easier to manage orders with Faire’s free SaaS platform, and to get access to net 60 payment terms and free returns.

When to try it

Clearly not everyone can take advantage of this channel. Can you? It primarily boils down to three factors:

  1. Does supply have enough of their own demand and enough leverage over them to convince them to transact via the marketplace?
  2. Is supply sufficiently incentivized to do that via better fulfillment, lower costs, better tools, better data?
  3. Is that demand promiscuous enough where they value the access to other suppliers once on the marketplace?

Businesses like Grubhub, Faire, and Eventbrite have all three qualities. Uber and Airbnb are examples that fail on the first point. Most peer-to-peer marketplaces like Poshmark fail on the second because suppliers would much rather transact off platform via something like Venmo, Paypal, or Cash app to save fees. Upwork and Thumbtack fail on the third because once a buyer has a supplier they trust, they tend to stick with them.

If your business meets all three criteria and you’ve been banging your head against a wall on SEO or paid acquisition, using supply to drive demand is the first thing you should try.

Just don’t over-rotate. If you’re not careful you can accidentally pivot a marketplace business with cross-side network effects into a SaaS-like network that does not have cross-side network effects. You will then find you struggle with slower growth, generally lower take rates, and a weaker relationship with the demand side of the marketplace.

How to pull it off

Here are three things to focus on to enable this channel at scale.

1/ Make it clearly better for suppliers to transact through your marketplace vs. through other marketplaces or directly with their customers. Examples:

  • Low or no fees for seller-referred transactions Examples: Faire Direct, Grubhub’s $1 fees for website ordering, Etsy Share & Save
  • Better payment terms than offline transactions receive e.g. get money faster or pay slower Examples: Faire net 60 terms
  • More efficient workflows when transactions are processed through the marketplaces that save the supplier time Examples: Eventbrite’s Mailchimp integration
  • Take on financial risk for these transactions suppliers would need to manage on their own if processed outside the marketplace. Examples: Turo’s car insurance, Bounce BounceShield

2/ Create a referral incentive for supply that is meaningful but still hits your payback thresholds. Notes:

  • Make sure to measure the value of seller-referred customers separately from other acquisition channels. They tend to be lower lifetime value due to the lower fees mentioned above and perhaps sub-par onboarding to the full marketplace value
  • If this incentive is shared with demand, fraud detection is necessary to maintain effective payback periods over time

3/ Create marketing tools that make suppliers better at attracting more transactions. Marketplaces are generally more sophisticated marketers than suppliers at scale, and suppliers know this. Examples:

  • Website builders or embedded checkouts into suppliers’ own websites that are better optimized for conversion. Examples: Eventbrite’s Embedded Checkout, Shopify Shop Pay
  • Optimize supppliers’ Google Local presence. Examples: Grubhub, Bounce
  • Email & SMS marketing tools. Examples: Eventbrite Email Campaigns, Zillow Contact Manager
  • Pooled performance marketing data from all customers on people most likely to convert and tooling to more easily target them. Examples: Etsy Offsite Ads, Eventbrite Boost

If you can do these three things, you will be well on your way to creating one of the most powerful demand acquisition channels for marketplaces.

Be sure to subscribe to my Substack to catch future posts.

Currently listening to Casey’s 2023 playlist.

A Framework for Developing Cultural Values

Every startup talks about their cultural values, but very few live them in a way that better allows them to win. Most are too vanilla and therefore forgotten by the team over time. There is a difference between cultural values, where companies say how they intend to operate on a spectrum compared to other companies, and principles that should apply to all companies trying to become great. Assuming good intent of co-workers, for example, is not a cultural value; it is a principle for great teamwork inside a company. It’s a dominant strategy for every company to operate that way. Cultural values should state the way this particular company does things that might be different from others. When cultural values work, they, along with dominant company building principles, should impact who gets hired, fired, and rewarded.

During the pandemic, Eventbrite felt like our company strategy had changed, we had a lot of new leadership, and our values weren’t helping us deliver against our new strategy. This was also a time where there was a lot of corporate pressure from employees on how they engaged on social issues, which was both generally distracting for a company trying to survive in war time. Other companies at the time were making headlines for, perhaps, some over-reactions to internal conflicts surrounding these topics (see Coinbase and Basecamp), and social media was over-reacting to those changes too.

So, we defined a process to evolve our cultural values, which I recommend all companies go through over time. I’m publishing the framework on how we did as a few of the startups I advise have found it helpful. This is just a framework; you don’t have to use it. The important part is to be intentional about your cultural values, and to keep them up-to-date with how you actually want your company to operate.

David, our Chief People Officer at the time, and I were inspired by a Harvard Business Review article called The Leader’s Guide to Corporate Culture. The article describes eight predominant cultural styles across companies. I created a spreadsheet that broke down these cultural styles based on a few different attributes, and asked our executive team to mark what they believed their predominant style to be. The attributes we focused on (which could be tweaked for your company) were:

  • Secretive vs. transparent: Do you default make things need to know or try to share everything?
  • For profit vs. for good: Are you a non-profit, willing to trade off some profit for your values, or entirely profit driven?
  • Independent vs. interdependent: Do you encourage people to move on their own or build consensus so everyone rows in the same direction?
  • Stability vs. flexibility: Do you try to keep the strategy the same or encourage pivots based on new information?
  • Whole self vs. focused on the work: Is work about work or are social causes a big part of the culture?
  • Top down vs. bottoms up: Do leaders make all the decisions or are decisions pushed down the org chart as much as possible?
  • Individual vs. team rewards: Do you reward everyone on the team the same or provide more rewards to those who deliver more impact?

You can view and copy the spreadsheet here for your own personal use. 

Hopefully, you can see on these attributes there is no right or wrong answer, just preferences of individuals on how they like to work. Obviously, I have my biases, but I have examples across every spectrum of companies operating differently and being successful. What’s important is the company is clear where on the spectrum they are. 

Each leader then picked for each attribute where on a scale of five points where they would ideally like to operate. For Secrecy vs. Transparency, for example, they would pick either: Very Secretive, Secretive, In the Middle, Transparent, Very Transparent. We then collectively ran an exercise where we defined what we believed the current leadership style of the company was today, and where the company operated today on these key questions. We then ran an exercise of where we wanted the company to go that might be different from how the company operates today.

As an executive team, we learned a lot about each other and the Eventbrite culture through this exercise. We unanimously agreed that Eventbrite was a Caring and Purpose culture as defined the HBR article. And we liked a lot of those attributes. We also noticed that a lot of the new leaders Julia had hired recently reflected more of a Learning and Results orientation. This seemed to be a deliberate exercise by Julia, if a bit of a subconscious one. So we wanted to communicate to the company more of an intentional move in that direction.

Once we aligned on these attributes and styles, what we did not do was write new cultural values. Because one of the attributes we wanted to change was to be more bottoms up, we then went to the rest of the company with this desired change, and formed a group of people across the company to develop the new values based on these directions. By workshopping every so often with David and myself, this working group defined new cultural values we then introduced to the entire company.

You will not come up with the same answers the Eventbrite executive team aligned on, or perhaps even debate the same attributes. But spending time on this alignment and using it to communicate clearly with the rest of the company can mitigate a lot of cultural issues as well as drive clarity on how certain decisions should be made beyond personal preferences.

Be sure to subscribe to my Substack to catch future posts.

Currently listening to my Synthpop playlist.

When and How to Build Second Products

This is part three in a series of posts related to some presentations I did for the TCV Engage Summit. The Summit gathered ~40 CPOs and product leaders to chat through topics centered around product development and product-led growth. This year, topics ranged broadly from incorporating AI to deliver world-class consumer experiences to defining and measuring different forms of community-powered growth. You can read parts 1 and 2 here and here.

In a previous post, I talked about how product work post-product/market fit shifts from zero to one innovation to features, growth, and scaling work. But a question founders and teams often ask is when do we start layering in innovation work again that creates new value props. In Reforge terms, we call this new product expansion. I recently did a talk for TCV’s Engage Summit where I explained the different types of product expansion, when to start building that second product with a new value prop, and how to know if it’s successful.

Why Second Products Matter So Much

Why do we even care about second products? Don’t some of the best companies in the world win with one dominant product? Well, increasingly that’s not the case. Companies can rarely ride one product into the IPO sunset anymore. Yes, the headlines are filled with many of these examples, such as Google in the 2000s or Zoom in the 2010s, but these examples reflect an environment that is becoming increasingly rare. The tech IPO narrative used to reflect stories that would include much of the below:

  • Large markets
  • Low or stagnant competition
  • Rapidly growing markets
  • Strong network effects or economies of scale
  • Scarce talent pools

A lot of those bullets can be explained by just the growth of the internet, and there being no entrenched internet-first competition. The maturity of the internet means most of these are no longer the case. Almost every recent tech IPO is multi-product at time of IPO, and the dynamics of their markets appear much different:

  • International competition
  • Multiple startups in the same space
  • Incumbents are tech native, no longer asleep, and copy what works from startups quickly
  • There is talent across a wide range of companies and skills
  • Network effects are no longer impenetrable

Uber, Instacart, Doordash, Unity, Klaviyo, Nubank, Toast and many other recent IPOs all reflect this new reality.

The Types of New Product Expansion

There are many ways for a company to expand its product offering, with different levels of difficulty. The main vectors on which product expansion should be evaluated is whether the expansion changes the product, changes the target market, or changes the core competencies required to deliver the product’s value. I highlight six different types of product expansion, in increasing levels of difficulty based on these vectors.

Expansion Type Product Market Core Competency Examples
Geography Existing New Existing Grubhub LA, Pinterest Brazil
Category Existing New Existing Whatnot Sneakers, Thumbtack Home
Format Existing Existing New Netflix Streaming, Snap Spectacles
New Value Prop New Existing Existing Uber Eats, Hubspot Sales
Platform Existing Mixed New Shopify App Store, Salesforce App Exchange
Strategic Diversification New New Existing AWS, Cash App

Geographic and category expansion skills are fairly well developed in software businesses. Companies build a deep understanding of how they achieved product/market fit in the first market, and make as few tweaks as possible to adapt the product/market fit to these adjacent audiences. Most marketplaces and social networks have executed these playbooks fairly well.

Format changes are usually only required around platform shifts already occurring or platform shifts a larger company is trying to drive. The last large one was mobile, and most internet companies were able to replicate their success in mobile.  Netflix and Snap have worked on more interesting format shifts, building on entirely new technologies to deliver their value props on new forms of media.

New value propositions are what we traditionally think of for second products, and will be the focus of the rest of this post. This is creating a new value proposition for your existing audience so that you can acquire, retain, and/or monetize them better. In extremely horizontal products, it may be wiser to launch a platform than build a lot of this new product value yourself, but this requires massive scale to attract external developers, and is very difficult to execute. I have written more about platforms here. Strategic diversification is a much rarer phenomenon where a core competency you have built internally is marketable for an entirely new value prop and audience, like Amazon leveraging its core ecommerce infrastructure to sell to other developers, or Square leveraging its financial expertise in SMBs to launch a consumer fintech product with Cash App.

New Product Value and S-Curve Sequencing

In my previous post, I talked about S-Curves. In that post, I mentioned that sequencing from an original S-Curve to a next S-Curve is the key to long term sustainable growth. That sequence can come from finding a new growth loop, but eventually that next S-Curve will require new product value to be created. This is what I realized when I joined Eventbrite. Eventbrite initially found this success with a content loop around event creation.

To continue its growth, instead of investing in new product value, Eventbrite kept grafting new growth loops onto this core loop to acquire more event creators and drive more ticket sales per event, creating a much more complicated growth model that looks like the below.


What became clear after building this model of how Eventbrite grows is that all of this effort would no longer drive the kind of growth Eventbrite needed to be successful on the public markets. We could no longer acquire event creators and ticket buyers fast enough, and we didn’t make enough money from them when we did. If we wanted to grow sustainably, we needed new products. And we probably needed them yesterday. It’s not that the company hadn’t ever tried to invest in new value props, but those that could create significant new growth had eluded them.

When To Invest in New Products

If you want to be proactive in thinking about when you should invest in building new products vs. diagnosing a growth problem and determining new product development requiring years of effort to be the solution, how do you do that? Well, the first step is tracking what impacts the need for a second product inside your company. Besides building a growth model and forecasting your growth from it, which I absolutely recommend you should do, what are the factors that contribute to how quickly you need to be investing in that second product after the first product finds product/market fit?

Historically, the factor that most people use as a heuristic is the business model. Traditionally consumer businesses have longer S-curves, so there is less of a need for a new product to drive growth. B2B requires suite expansion. Why does B2B require suite expansion? Well, they usually do not have network effects which makes marginal growth harder, of course, but the main reason is competition from bundled competitors. Acquisition, retention, and monetization potential of your first product is another reason B2B tends to expand earlier. Second products influence sales efficiency and profitability dramatically. Next is the size and growth of the market. If the market is large, your product can grow inside it for a long time. And if the market is growing fast, market growth can frequently drive enough company growth on its own, like, say, Shopify with ecommerce. The smaller the market is, the faster you need to expand the addressable market to grow. The last factor is how natural product adjacencies are for your first product. Generally, in B2B, product adjacencies are more obvious and less of a gamble to invest in. Launching successful consumer products is very hard with a very high failure rate.

But I’m going to show you why if you pay close attention to these other factors, the business model can be a red herring.

New Product Expansion by Business Model

Let’s break some examples down by business model and start with pure consumer businesses. Pinterest and Snapchat were compared a lot because we started scaling at around the same time. And even though they are both the same business model, you can see some of their attributes look a lot different. 

First, no one actively competed with Pinterest during its rise to be the primary way people discovered new content related to their interests. Snap, meanwhile faced an aggressive competitive response from Instagram as they grew to be a place where friends interacted around pictures. From a customer acquisition perspective, the companies grew in very different ways too. Pinterest grew by capturing users searching for things related to their interests on Google while Snapchat grew virally. Their retention strategies were also different. Pinterest primarily increased engagement by learning more about what you liked and recommending content better and better matched to your interests over time. This is usually a strong retention loop. Snapchat built out your friend graph, but didn’t really get much stronger after that. In fact, too many friends might be off putting. The most important difference was the monetization potential. Pinterest’s feed of content related to your interests is a perfect model for integrating advertising and commerce, the two best consumer business models. Disappearing photos however was not a good fit for either of those models, and likely best lent itself to subscriptions and virtual goods, both largely unproven at consumer internet scale. Lastly, Pinterest grew adjacencies by making the product work better with interests in different local geos and in different categories e.g. travel vs. fashion. Snap had similar geographic growth, but had some additional format and product adjacencies.

Okay, so let’s look at how Pinterest and Snapchat grew their product offering over time. We’ll focus on the consumer, not advertiser side of the equation for this example, though obviously both companies built advertising products. The companies launched around the same time, and launched their second products around the same time. Pinterest significantly evolved how its core product worked, changing both the acquisition and retention loops over time. Acquisition shifted from a content loop built on top of Facebook’s open graph to a content loop built on top of Google with SEO. The way Pinterest retained users also changed, from seeing what your friends were saving to getting recommended the best content related to your interests regardless of who saved it. Snapchat did not evolve their core product nearly as much. But Snapchat’s second product was a lot more successful. Snapchat Stories was a huge hit. Pinterest around the same time released Place Pins, a map based product that did not find product/market fit and was deprecated. Both companies also launched additional new products in the coming years. Snapchat found new product success again with Discover, and Pinterest again failed building a Q&A product around saved content.

So wait a minute? You’re telling me Snap succeeded multiple times in product expansion where Pinterest failed, yet the companies are valued at about the same. What gives? Well, it turns out Pinterest didn’t need to expand into new products because its initial product had great acquisition, retention, and monetization potential, albeit with some evolution on how they worked. Iterating on its initial value prop was the unlock, not creating new value props. In fact, the new product expansion work outside of expanding countries and categories was a distraction that probably prevented the core product from growing faster. The company might be worth double if it had not spent so much time trying to develop new products. Snap, however, probably would not have survived without product innovation because its first product had low monetization potential. They needed those new products to work, and they did.

Let’s look at an example in SaaS. I had the pleasure of working with both Figma and Canva as they were developing. I was an advisor to Canva starting in 2017, and got to work with Figma while I was a growth advisor at Greylock, which led the series A investment. It’s a fascinating example of two design tools targeting entirely different audiences, basically designers and non-designers. 

At the time of their launch, Figma was in a competitive space with legacy products from Adobe, and many tech companies were using Sketch. Theoretically, Adobe’s Photoshop was the competitor for Canva, but it was much too complicated for laypeople to use, and much of Canva’s pitch was that it was Photoshop “for the rest of us”. Both could acquire users by having creators share their designs. Figma looked like it would be a much higher retention product as it was multi-player from the start, and applicable to larger businesses. Canva was more of a single player and SMB tool. As a result of this, it looked like Figma would monetize a lot better, with a classic per seat model selling to enterprises, with Canva having a lot of single user subscriptions. 

At the time, investors didn’t think the design market was one of the larger markets out there (they were wrong), but everyone did think the category was high growth. Both companies had some nice theoretical adjacencies in terms of formats they could work in, new products they could create, and platform potential.

Both companies evolved how they acquired users over time, layering in sales, and Canva got a huge boost from SEO. Both companies also evolved their retention strategies. Figma became a tool not just designers to collaborate, but for those designers to collaborate with their peers in engineering and product. Canva created lots of ways for users to not start from scratch with community-provided templates and stock photos to leverage. 

Figma launched its first new product in 2019, called Figma Community. It intended to create  a Github-like product for designers, or perhaps a Dribbble competitor. It has not reached the company’s expectations. Canva launched Presentations in 2021, and it has become a heavily used product. Both companies have continued to invest in delivering new value props. Figma launched Figjam, a Miro competitor in 2021. It has not become the Miro killer the company imagined as of yet. Canva launched its video product also in 2021, and it continues to gain traction, along with a suite of other enterprise bundle features more recently, like a document editor.

So on paper, it looks like Figma’s in a competitive space. Canva is not in a competitive space. But while Figma has had mediocre product expansion and is still being sold for potentially $20 billion, Canva is grinding on building out a suite to succeed. Why? Because Figma’s product/market fit so quickly surpassed what was on the market that products ceased to become competitive over time. And while Canva didn’t have competitors, it became a substitute to the big incumbents Adobe and Microsoft, forcing them to build copycats and respond. Canva as a point solution likely loses out to their bundles if they don’t expand their suite successfully.

Time and time again, we see two things. One, companies in the same space may need to think about new value prop development much earlier than other seemingly similar companies. Two, we see new products not inflect company growth when they are a random bet on innovation. New products tend to work when they have to work for the company to succeed. If you can still grow your core product, they rarely get the focus they need to succeed, and furthermore, might be a less efficient use of resources than continuing to grow the core business. Figma will not live or die on the success or failure of Figjam. Canva might need these new products to be successful to stay competitive long-term. 

Portfolio approaches that recommend some percentage of development on innovation vs. features vs. growth vs. scaling tend not to be where massively successful second products come from. Understanding your growth model, and betting big on new product development when you sense the company needs it, tends to be the more successful approach.

Okay, let’s look at a marketplace example. Here is where you see how marketplace strategy has needed to evolve over time. Older marketplaces like Grubhub were extremely profitable because they did not facilitate the transaction beyond payments. More recent startups like Instacart have needed to manage a significant component of the delivery of the value prop, which means its monetization potential out of the gate is much worse. 

Similar to Snap vs. Pinterest, Grubhub’s initial market was so large and so profitable that all new product value expansion did was limit the potential of the core product. Pickup cannibalized search results and lowered activation rates, especially in some key markets like LA that allowed upstarts to gain traction, notably Postmates.

Instacart’s initial product however required so many complex operations that it found it could not eke out real profits while paying groceries and pickers. But it could expand its network to CPG advertisers, replicating grocery market slotting fees in a digital product. So these companies had very different paths to similar market caps despite both being labeled marketplaces.

Last, but not least, let’s look at a consumer subscription example. Duolingo and Calm launched around the same time as consumer subscription apps. Both are in competitive spaces that struggle with retention because building new habits is hard for consumers. The language market is however considerably larger than meditation.

Both companies evolved their acquisition strategy over time, but Duolingo got a lot more leverage out of virality, keeping their acquisition costs much lower. Duolingo’s core product experience also got stronger over time through both data and manual improvement in lessons from user engagement, and laying in gamification tactics. Calm moved from web to app, and built in some daily habits that helped retention. 

What made Calm a much more interesting business though was the launch of Sleep Stories. Not only does expanding into sleep expand the target audience dramatically, it makes it easier for Calm to become attached to a durable habit. People have to sleep; they don’t have to meditate. Calm also was able to expand into B2B by selling Calm as a mental health benefit. Duolingo did not have the same success in new product expansion. While the core product continued to get better at covering more languages, new product efforts failed to create value, such as TinyCards in 2017. Yet, even with this fact, Duolingo appears to be a lot more successful than Calm, likely primarily due to the acquisition strategy and larger initial target market.

In these scenarios, it is not good to assume you are one of these companies on the left side of the table where your initial product/market fit will have such a large addressable market and lack of competition that you can scale successfully without new product development. It is also dangerous to assume you will need a lot of new product innovation when your initial target market ends up being quite large. What I urge companies to do is dig deeper into the attributes in these tables for their company, and re-ask these questions every year as we have seen many of these market dynamics shift dramatically over time.

How to Know If New Products Are Successful

So when is a new product “successful”? Well, the answer, surprisingly, is not product/market fit. If you’ve read some of my work, you know I define product/market fit as satisfaction, normally measured by a healthy retention curve, that is through its own engagement or monetization able to create sustainable growth in new users for a significant period of time.

But second products don’t need to do all of that to matter. Whereas a new startup isn’t going anywhere unless it figures out acquisition and retention (and maybe even today monetization), new products may only need to influence one of the three to be successful. But the key is, they need to influence it for the overall company, not just the product itself. So if a second product has high retention and can effectively acquire new users, but can never inflect the growth of the overall business, it’s not successful. 

This is why developing a growth model above becomes so important. It can tell you if the new product is developing fast enough to inflect growth of the overall business, and when that might happen. And if it isn’t, you can understand what it will take for that to happen. This is something that confuses product teams that work on new products inside larger companies. By the frameworks they understand, the new product “is working.” It has product/market fit, it’s growing, etc., but it can never grow enough to really help the overall company.

Most companies struggle to understand when they need to start investing in adding new product value vs. just continuing to grow off the traction of their initial product/market fit. But it is becoming necessary earlier and earlier in a company’s lifecycle due to a confluence of factors. In order for us to get better at building great, enduring businesses, we need to talk about the types of expansions that matter for companies, and assess at an individual company level what is required for the new phase of growth. Modeling your growth really is a helpful start, and digging deep into understanding the competitive landscape, the acquisition, retention, and monetization potential of your current business, the size and growth of your market, and what your natural adjacencies is becoming critical to make the right calls at the right time regarding new product investment. New products work when they have to. It’s time to ditch outdated portfolio practices and innovation teams, and build modern approaches around when to start building, investing hard in building new products when it is the right time, and evaluating their success or failure properly.

Currently listening to my Early Dubstep playlist.

On Platform Shifts and AI

At TCV’s Engage Summit in 2022, I gave a take on finding your next wave of growth, which you can read here. Sam Shank, founder and CEO of HotelTonight, asked me the question that everyone asks growth people, “What new channels are you seeing new consumer companies take advantage of?” My answer was disappointing as it always is, “What new consumer companies? Discord is the last one I have seen grow organically for a long period of time. Tiktok spent many billions of dollars buying up every ad they could on Facebook, Snap, and Google properties. It’s not really replicable.” I then proceeded to explain that consumer companies tend to arrive in droves during platform shifts, and we haven’t had one since mobile. But AI could be coming (editor’s note: it did). Sam quickly pointed out that while AI is a potential technological platform shift, it is not a distribution platform shift. And it’s distribution platform shifts that create new consumer opportunities. I’ve thought about this conversation a lot, and think I have a better framework to both describe what Sam was describing, and what that means now that the technological platform shift clearly arrived when ChatGPT came out.

What separates a major platform shift from a minor platform shift is a platform shift that enables both a technological shift (new ways of making things possible) paired with a distribution shift (new ways of reaching people with it). The internet and mobile both created new technological and distribution shifts that enabled lots of new multi-billion dollar businesses to be created, whereas “cloud” as an example made new things possible without any new distribution (favors b2b innovation) and crypto arguably enabled new forms of distribution (tokens), but didn’t fundamentally make many new things possible with technology. So nothing can stand the test of time in that space. I’d argue the only companies that have found product/market fit in crypto are companies that either enable or catch grifters. A more pithy way of saying this is the crypto space has created more criminal convictions than companies with product/market fit. Other things VCs have hyped in the past as potential platform shifts have largely neither made interesting new things possible nor driven new distribution opportunities (NFC, VR, Internet of Things, et al.).

What I realized having gone through the internet and mobile platform shifts is that the technological and distribution shifts did not happen at the same time. Platform shifts that create both technological and distribution opportunities happen in a sequence, not all at once. The internet created websites, but the search engine wouldn’t come along until later to become the dominant form of distribution. Mobile created mobile apps, but it was Facebook mobile ads, not the App Store, that became the dominant form of distribution for mobile apps. So, AI has come out and definitely created a technological shift that enables new ways to solve problems that couldn’t be done before. But AI lacks a new distribution channel. ChatGPT is “not it”, as the kids would say. At least not yet.

So, today, that means the traditional distribution methods need to carry the distribution of AI innovation. This favors either: 

  • incumbents who already have distribution
  • startups that can leverage traditional channels such as sales, virality, user generated content, or paid acquisition because their product value is deeply innovative and very marketable

But, this may not be forever. As I mentioned before, we shouldn’t really expect new distribution shifts to have happened yet. The App Store launched in 2008, and even though there was fervor around discovering apps on the App Store for a while with the “there’s an app for that” campaigns, that fervor died as did most of the apps featured. It was when Facebook launched mobile ads four years later in 2012 that apps exploded into multi-billion dollar companies. This is similar to the internet. People started getting online around 1994. Google didn’t come out until 1998. Sure, there were search engines before that (Lycos, Yahoo!), but they lacked the predictable distribution of Google. Word of mouth can’t scale technological shifts alone. They need scalable distribution methods, and usually new ones that take time to become obvious.

So, as an operator, this feels like 1997 or 2008. The Google and Facebook mobile ads of AI haven’t come out yet. Most of the companies that exist will die in the next five years like the internet bubble as they lack sustainable business models and distribution, but there are a few that won’t (Amazon, Ebay, OpenTable et al. survived the internet bubble), and much of the next gen after this wave will become very large. And unlike the internet bubble, incumbents are on top of it and many will do quite well capitalizing on this shift. Some will get destroyed, of course.

We also can’t bet on a new distribution channel coming for AI though. With every generation, companies that reach massive scale have gotten more efficient at preventing other companies from growing on top of them, at least for free. Google created a scalable way for companies to grow both organically with SEO and by paying for it with Adwords, and it still works decades later. Facebook, after flirting with a similar strategy to Google, decided to charge companies for all distribution on its platform. So, if this distribution channel never materializes, expect the impact of AI at consumer scale to be mostly coming from consumer companies that already have consumer scale vs. a bunch of new Facebooks and Googles. I’m rooting for those new distribution angles myself though.

Currently listening to my Future Bass playlist.

Finding the Next Wave of Growth: S-Curves and Product Sequencing

I’ve had the pleasure of speaking at TCV’s Engage Summit the past two years. The Summit gathered ~40 CPOs and product leaders to chat through topics centered around product development and product-led growth. This year, topics ranged broadly from incorporating AI to deliver world-class consumer experiences to defining and measuring different forms of community-powered growth. I never posted my talk from last year, so I’m adapting it into a blog post here, and will do the same for this year’s talk in the following weeks and as well as some follow up questions from the Summit I’ve had a chance to ruminate on.

Also, I publish on Substack now! So subscribe here.

After product/market fit, most companies’ obsession is not thinking about how to create their next amazing product. Their obsession is thinking about growth. Specifically, how do I get this product I know is valuable in the hands of everyone it can be valuable to. Most companies have a primary acquisition loop that drives this scalable growth, and unfortunately, there aren’t that many acquisition loops that really scale. Even when they scale, they eventually asymptote, and companies need to find new ways to grow. This can be new growth loops for the same product, or entirely new products. In this post, I’ll explain how to think about the timing of that, and show some of the successes and failures of my career.

As I have discussed in previous essays, product/market fit can be hard to interpret at the time. When you find product/market fit, problems don’t go away. Customers don’t stop complaining. In fact, they complain more, because they like the product enough to care. What they stop doing is leaving. And you start being able to acquire more of them in a scalable way i.e. an acquisition loop.

Because of this and other factors, when you find product/market fit, you can’t stop iterating. Product/market fit has a positive slope. If you find product/market fit and don’t continue to make the product better, a rising competitive landscape and customer expectations can have you fall out of product/market fit over time. But when you do achieve product/market fit, while you don’t stop iterating, your portfolio of what you work on needs to change.

At Reforge, we talk about the four types of product work. Once you find product/market fit, zero to one product/market fit work goes away entirely for a while as your portfolio shifts to different types of product work:

  • Features: improve current product/market fit
  • Growth: connecting more people to existing product/market fit
  • Scaling: being able to scale the product to more users and more teams internally
  • Product/Market Fit Expansion: new segments, markets, and eventually products


The growth work in particular becomes a major focus, strengthening and discovering new acquisition and engagement loops. Most companies when they find product/market fit with their first product only have one acquisition and engagement loop that is successful, and the job of most of the team is to refine and scale those loops. At Pinterest, I was originally in charge of building a new acquisition loop built on top of Google. It looked like this:

We eventually re-architected our engagement loops to be based around personalization instead of around friends. When you stitch these acquisition loops and engagement loops together, it creates a more complicated growth model that looks like this:

The acquisition loop now feeds new users into a personalization loop that increases engagement over time, and emails and notifications reinforce that loop by distributing relevant content to users outside the product to bring them back. The entirety of Pinterest for the first few years I was there was tuning these loops in one way or another. Eventually, the company needed to layer in new advertiser focused loops to monetize, but I’ll skip that detail for now.

When I arrived at Eventbrite, the company was a lot more mature than when I started at Pinterest. But similar to Pinterest, it started with one acquisition and engagement loop driving its growth.

Creators market their events to bring in new ticket buyers. Many of those ticket buyers, once introduced to Eventbrite, start creating events themselves. And when event creators are successful at selling tickets, they come back and create more events. But Eventbrite didn’t stop there. It kept investing in making its overall growth model stronger.

Why did they do this? Well, all growth loops eventually asymptote. If you get good at modeling your loops, which basically takes the diagrams above and turns them into spreadsheet based forecasts of the impact to your business, you can start to predict when they will stop driving the growth the company needs. Modeling both helps you predict when those asymptotes will happen and unconstrain those loops by finding their bottlenecks and alleviating them. At Pinterest, we 5x’d conversion rate into signup over time, and doubled the activation rate of signups to engaged users over time as a couple of examples.

Some constraints in your growth loops can’t be fundamentally unconstrained by optimization though. The company requires either new growth loops or new products to acquire, retain, or monetize better. Modeling your loops helps you start investing in building out those new growth loops or products well in advance of when you need them to sustain your growth, because of course developing them takes much more time than improving a current loop. We think of this as sequencing different S-curves of growth.

By my arrival as an advisor by 2017 and CPO by 2019 at Eventbrite, the company had layered on many more acquisition loops onto its original loop to continue to grow, creating a much more complicated growth model.

Now, I know this looks complicated, but all that is really going on here is Eventbrite took its monetization of ticket sales and re-invested all of that money into new acquisition loops to bring in more event creators (sales, paid acquisition, content marketing). Also, Eventbrite took the increasing scale of event inventory created on the platform and started distributing it themselves to drive more ticket sales per event to places like Google, Facebook, Spotify, and its existing base of millions of people who had bought tickets to previous events.

People don’t talk enough about how much S-curve sequencing work went on at all these successful companies, so I wanted to give you a taste of what it looked like across my experience at Grubhub, Pinterest, and Eventbrite because it’s a lot, and a lot of it didn’t work. Let’s start with Grubhub summarizing ten years of decisions that both helped and hurt Grubhub as it scaled to be a public company (+’s show up where I think the decision helped, and -’s where I think the decision hurt):

  • 2004: Grubhub co-founder collects menus of Chicago neighborhood restaurants, scans them, and puts them online (+)
  • 2005: Grubhub expands to cover all of Chicago (+)
  • 2006: Grubhub launches online ordering from restaurants (+)
  • 2007: Grubhub optimizes sales model and expands into second market (+)
  • 2008: Grubhub unlocks demand side channels and refines expansion playbook (+)
    • Grubhub launches Boston and New York (+)
    • Grubhub landing pages for restaurants that deliver to X start ranking well on Google (+)
    • Grubhub unlocks paid acquisition to drive demand (+)
  • 2009: Grubhub scales market launch playbook (+)
    • Grubhub switches from flat fee to percentage model (+)
  • 2010: Grubhub launches pickup (-)
    • It doesn’t find product/market fit and hurts delivery use cases (-)
    • Grubhub now launching at least one new market per month (+)
  • 2011: Grubhub acquires Campusfood and launches restaurant websites (+)
    • Grubhub acquires Campusfood to expand to many college markets (+)
    • Grubhub acquires Fango to build in-restaurant tech (+)
    • Grubhub launches restaurant websites to drive in-restaurant growth (+)
  • 2013: Grubhub acquires Seamless (+)
  • 2014: Grubhub goes public and starts building a delivery network to compete with Uber and Doordash (+)
    • It doesn’t matter as those companies raise billions of dollars to destroy Grubhub’s network effect (-)

What you can see here is despite a successful outcome of an IPO and $7.6 billion exit, Grubhub made a lot of mistakes. If you strip those mistakes out, the sequencing of S-curves looks like:


The main lessons that matter here to me are that Grubhub tried product expansion too early with pickup. But market expansion became a major strength and well oiled machine through sales and SEO expertise as well as strategic M&A. That strategic M&A failed them, however, in responding to the threat of delivery networks. Grubhub was integrating its largest acquisition when Doordash and Uber Eats rose to prominence, and while Grubhub acquired over a dozen companies, it never acquired the one that was truly disruptive (Doordash).

Okay, let’s do Pinterest in the same format:

  • 2010-2011: Founder visits DIY/Craft Meetups and convinces Influencers to start “Pin It Forward” Campaign (+)
    • This gets people to learn how to use the “Pin It” functionality in their browser (+)
    • Pinterest uses Facebook Sign-In to bootstrap network of friends as more people join the platform (+)
  • 2011-2012: Pinterest leverages Facebook Open Graph to share every Pin into users’ Facebook feeds (+)
    • Pinterest starts to amass enough content to make discovery, not saving, primary value prop (+)
    • Retention and frequency of use improve (+)
  • 2013: Facebook turns off Open Graph and growth stops (-)
  • 2014: Pinterest fails to unlock growth with new products, but does unlock User Generated Content distributed through SEO (+)
    • Pinterest launched a maps product, a Q&A product, and a messaging product, and all fail to drive growth (-)
    • Pinterest finds another channel in Google to distribute its high quality content to after Open Graph turned off by Facebook (+)
    • Users come in with less match to existing network, so friend graph ceases to drive ongoing discovery. Retention decreases. (-)
  • 2015: Data network effects kick in (+)
    • While friend graph ceases to work, Pinterest now has the scale of content to recommend great content just based on users’ interests. Moves to interest, not friend based discovery. Retention improves again. (+)
    • Pinterest pauses all U.S. work to make sure we unlock international markets (+)
    • Pinterest tries to re-ignite user sharing and fails (-)
  • 2016: Pinterest crosses 50% international active users (+)
    • Focus shifts to building advertising business to make money (+)
    • Growth team starts seriously experimenting with paid acquisition as new channel (+)

Despite Pinterest being worth *check’s today’s stock price* $21.5 billion on the public markets today, you still see a lot of the mistakes we made. Too much new product development that didn’t pan out and too much trying to regain what we had lost vs. leaning into new areas that were working. Network effect products rely less on new product innovation unless it’s the only way to monetize. And Pinterest tried the harder expansion before the easier ones. Market and category expansion tend to be much easier than product value expansion. But, Pinterest did make a very successful pivot from direct network effects to data network effects and from Facebook to Google as the primary distribution channel. When you strip the failures out, our success looks like the following sequence:

Okay, for the last one, let’s do Eventbrite:

  • 2006: Eventbrite launches to allow event creators to accept payments online (+)
  • 2007: Event creators start putting $0 in the payment field to create free tickets, driving huge awareness (+)
  • 2008-2012: Eventbrite builds more features to help event creators run their business and includes them in ticket fee (-)
  • 2012: Eventbrite builds sales team to scale to more upmarket event creators (-)
    • Eventbrite launches new countries with sales-led strategy (-)
    • These countries never build the self-serve growth motion of the U.S.
  • 2016: Eventbrite launches consumer destination to help consumers find events (+)
    • SEO landing pages featuring events in different cities become large drivers of ticket sales (+)
    • Eventbrite begins scaling emails to consumers of events they might be interested in (+)
  • 2017: Eventbrite launches packages and acquires Ticketfly to move upmarket into the enterprise music segment (-)
    • Packages makes Eventbrite more money in the short turn, but drive churn and less acquisition over time (-)
    • Many Ticketfly customers are a poor fit for Eventbrite from a service / functionality perspective. Segment is low growth. (-)
  • 2018: Eventbrite acquires Picatic to build developer platform (-)
  • 2020: Pandemic hits, and Eventbrite rewrites strategy to focus on independent, frequent creators and help them grow
    • Focus on self-service and helping creators drive demand
    • Cancel separate music product and developer platform
  • 2021: Eventbrite launches Eventbrite Boost, a suite of tools to help creators improve their own marketing
  • 2022: Eventbrite launches Eventbrite Ads to help event creators reach more consumers searching for events on Eventbrite

Since this shift is happening in real time, I’ll describe the S-Curve sequencing Eventbrite was investing in as of the end of my full-time role. The value prop is shifting from payments and ticketing to helping event creators grow their ticket sales. Eventbrite has launched new pricing with tools like marketing tools that help event creators get better at their own email and performance marketing as well as let them get more distribution inside Eventbrite’s platform. The revenue from this will help drive more investment in the consumer product side of Eventbrite, which hopefully drives more consumers looking to Eventbrite to find things to do and buying more tickets from our creators.


Hopefully you see from these examples that sequencing S-Curves to drive growth of companies over the long term is not only quite difficult, but the craft of doing it is under-developed. All three of these companies made some critical successful moves as well as major mistakes that set them back years. I hope that by studying these and other examples startups can get smarter about how they sequence their S-Curves and drive long term success for their companies. In my next two posts, I’ll go deeper on how to think about how platform shifts like AI affect this and publish a lot more on when and how to invest in building your second product successfully.

Currently listening to my Rhythym & Bass playlist.

And don’t forget to get on my Substack list for future posts here.

Why Consumer Subscription Is So Hard, and What to Do About It

I was recently on Lenny Rachitksy’s podcast again, and one of the topics we discussed was consumer subscription business. I thought I would actually write down a lot of those thoughts and add some more depth for founders and employees working on these businesses. Don’t worry I got more marketplace content on the way as well 🙂

Shortly after I got into tech, investors started to fall in love with subscription business models, mostly on the B2B side. Across many different problems, subscription software sold over the internet seemed to produce dominant tech companies left and right. Even incumbents like Adobe and Microsoft rebuilt their businesses around these subscription models to unlock new growth. And unlike most trends in tech that start on the consumer side and then migrate into B2B over time, the subscription model actually did the reverse. So years after Salesforce, Shopify, etc. became behemoths, people started adapting consumer software to subscription models. Founders and venture capitalists would preach the gospel of predictable revenue and sustainable growth, powered by growth of apps sold over the App Store. These companies have largely under-performed their B2B counterparts, and I’ll use this post to explain why and how to produce superior returns in this category.

Why is B2B Subscription So Good?

While the pessimist would argue B2B subscriptions business actually aren’t that good either and more so looked good during a zero interest environment where “seat growth” was a given for most of your customers, I believe B2B Subscription has some durable advantages we will see play out even in the weakest of markets compared to many other business models. The average VC tweet storm or LinkedIn thinkfluencer will tell you it’s all about predictable revenue. And that’s kind of at too high a level to be instructive in my opinion. First off, revenue isn’t that predictable, but what does improve predictability is that your customers are more rational in their decisions. You tend to be able to evaluate which potential customers will become good sources of revenue, and won’t be terribly surprised by their usage of a product, whether they go out of business, will they grow their usage / seats / plans, etc. And most of these customers are sizable enough where the revenue from them creates very sustainable growth levers whether it be sales-, product-, or marketing-led growth, sometimes a combination of the three. But more importantly, when customers do grow, they create a phenomenon known as net dollar retention. So in any B2B subscription business, sure, some customers will churn. But the ones that do retain tend to invest more in your product over time, growing revenue per customer in a way that covers more than the churn of other customers. Many B2B subscriptions companies are seeing their first years of seats / usage contract due to the economy, but in general, outside of some bad years, net dollar retention will be a core feature of their growth model. 

What is different in Consumer Subscription? In short, everything. 

1. Churn will be higher. Average revenue per customer will be lower. Net dollar retention non-existent.

Consumers are not rational. It is very hard to predict who will churn vs. build habits around your product. So, on average, churn in consumer subscription businesses tend to be a lot higher. Also, the amount made per customer tends to be a lot lower as consumers have less spending power than businesses on average. What makes this fact even worse is even when you activate consumers effectively into paying users who build habits, you do not increase the amount of money you make from them over time outside of increasing prices. This means net dollar retention, the best feature of a B2B subscription growth model, is not even present in a consumer subscription growth model.

When you look at the consumer subscription products that have done the best at scale, they have won by spending an incredible amount of spend on content to prevent churn. These are companies like Amazon Prime, Netflix, and Spotify. That amount of spend on content is usually not replicable for startups.

2. Payments will be less optimized and more expensive

On top of this, mobile apps have become the most dominant product experience for consumer experiences, and the app stores take a significant percentage of subscriptions purchased through them, and prevent alternate forms of payment to avoid that tax. So the margin structure of these consumer businesses are significantly hampered compared to their B2B counterparts. To make matters worse, these app stores are also much worse at collecting payments. One way B2B growth teams help their companies improve is addressing involuntary churn through payment failures. Not only is Apple not near as good at this as, say, Stripe, Adyen, et al., because it controls the entire experience, your growth team cannot optimize it to improve payment failure rate at all. Fortunately, the lockdown on in-app payments is starting to break, allowing consumer companies to improve payment conversion on mobile web while improving margins at the same time, but it’s been a pretty annoying drag on an already difficult business model.

3. Customer acquisition is much harder, less scalable, and has fewer options

B2B subscriptions are sold to teams and companies at scale. Consumer subscriptions are sold to individuals and, at best, families. We already know that means the average order value is that lower. But what does that mean for your acquisition loops. Well, first off, sales is out of the question due its cost vs. the return of a subscription. Secondly, some of the viral and content loops that you see in B2B subscription like sharing a file, inviting a co-worker to collaborate / chat, etc. aren’t really possible in these consumer products. This mainly leaves paid acquisition as the lever for growth. You probably already know my position on paid acquisition as the main lever of growth, but let’s cover it quickly again.

Every company uses paid acquisition to target its best potential customers first. And this usually works with healthy payback periods. But, to scale, the company needs to target more and more customers who look less like the core customer over time. They respond to the ads less, convert worse on the landing page into trials, convert from trials to subscriptions worse, and retain worse after subscribing. All of this leads to predictable degradation in payback periods year over year until, eventually, you run out of people to acquire profitably and paid acquisition is no longer viable. Oh, and by the way, while this was going on Apple decided to torpedo all the ways you track effectiveness of paid acquisition too.

When we were evaluating investing the series A for Calm at Greylock, this is what spooked us. At the time, Calm had only decent annual retention. We could model out a time in the future when it would be impossible for them to grow based on their current numbers. Now, in Calm’s case, they are one of the few companies to improve their annual retention over time, which we’ll talk about in solutions to these issues below.

How to Solve the Systemic Challenges of Consumer Subscription

Now, I’m not going to just doom loop you and this post after ranting about how bad a business model consumer subscription is. I’m on the board of a consumer subscription company after all. Clearly, I believe there are solutions to these problems. Let’s talk through the best ways to fight these limitations and still create big outcomes.

1. Leverage network effects to solve retention and acquisition issues

My main selling points on network effects is that they allow your core product experience to get better faster than the customers you acquire get worse. The next generation of consumer subscription businesses tend to create product experiences that get better over time through either leveraging the data of their users or by offloading content creation costs to suppliers. Duolingo has done a masterful job of keeping retention high in a space with normally high churn because their lessons get better every day based on the feedback loop of their customer usage. We call that a data network effect. Spotify has exponentially more types of content (music, podcasts, video) and exponentially more artists than when it originally launched which attract more listeners. That’s a cross-side network effect.

Companies like Beek and Fable as startups continue to add new creators of content that improve selection or discovery of content for consumers, which keep them subscribed. But also, those creators make money based on overall subscription revenue of the app, so they become promoters of the app and a significant source of low cost acquisition. This allows these companies to rely a lot less on paid acquisition, and some don’t even have it as part of the mix at all. I also think there are many more opportunities to create multiplayer consumer experiences (or direct network effects) to drive low cost acquisition and better subscription retention because your friends or family keep pulling you back into the app. We mostly see this in the games category today.

2. Go multi-product earlier in your lifecycle to make the product stickier and raise price

It’s very hard for single product solutions to maintain the type of long term retention without network effects. So, if network effects do not make sense for the type of product value you’re delivering, launching new products to monetize your existing customers better and open up new customer segments can ease the burden on customer acquisition, raise monetization rates, and raise retention all at the same time. Calm was able to scale out its sleep stories product in a way that raised the retention rate of its meditation customer base as well as open up segments that were less interested in meditation. It turns out selling a product solution for something people have to do everyday (sleep) has a much bigger market than a habit a small percent of the world does (meditation).

3. Open up less saturated acquisition channels

Most consumer subscription business treat their content as their proprietary secret sauce and keep it under lock and key inside paid subscriptions. Much of the time, this means that content isn’t doing all it can to attract new customers who are not even aware of your product. Masterclass is a great example of re-using a lot of the amazing content they sell in their courses and repackaging it for search engines as a taste of what the full courses offer. This has allowed a company that historically 100% grew via paid acquisition to diversify its acquisition sources, bring down payback periods, and find new audiences. Spotify’s playlist sharing was a key growth driver in its early days as lists were shared among friends and publicly on the internet. Spotify and Hulu have also bundled their subscription model to find new audiences and to improve retention for both products.

Thinking about new platforms as well as channels works here too. Most subscription businesses start as apps, but the web opens up a new acquisition channel with significantly better margins because you don’t have to pay the in-app purchase tax. Many of the companies I have worked with have found ways to get conversion on the mobile web just as high as in app over time, or at least use annual subscriptions to dramatically improve payback periods.

4. Start selling to businesses (you knew this was coming, right?)

Well, this one is obvious, and basically the same suggestion as #3. Creating a B2B offering allows you to target a new customer business with a new acquisition loop in sales that can acquire hundreds to thousands of people at the same time inside companies. Headspace and Calm have done a good job of expanding into this model as an expansion from their consumer roots.


I want to be very clear. These suggestions will not necessarily be a panacea, and they may not make a meaningful enough improvement to take your consumer subscription business beyond scale. Many of the companies I mentioned still may not have long-term success. This is why I advise founders to think through these challenges and opportunities during the zero to one phase of building their vs. being surprised how hard things will be in the growth stage. This is a very hard business to scale toward a venture outcome, and you want every possible thing you can to be working for you because so much of the default business model works against you.

Currently listening to my Avant Pop playlist on Spotify.