Chapter 2: Distillable
A Promising Start
When I started building Distillable (formerly known as Dolphi), I started building an online learning platform. I saw all these advisors and experts in the startup and consulting market and thought, “Hey, wouldn’t it be cool if I could democratize access to them?”
I had already published a book, and the book felt like the beginning of an early success. The feedback was overwhelmingly positive, and it opened doors to press, corporate engagement, fireside chat, and valuable connections. I started a podcast to interview other leaders and ran cohorts related to the book. It felt like the beginning of something big.
Alongside the book launch, I was running writing circles with tech leaders who often sought my advice on publishing online. I assumed that because I had a successful experience, I could apply the same approach to help other people productize their knowledge too.
I was also quick to connect with strong mentors, which made the early stage seem promising. When I started networking and sharing ideas, I often received responses like, “Oh, that sounds cool.” I mistook this for validation. In hindsight, these reactions were more like automatic responses. They didn’t offer any substantive feedback about our product or idea. What interested people was how we secured those high-profile connections. This led to a big misunderstanding about what our early adopters truly valued.
The (Underestimated) Nike Effect
At the same time, the creator economy was approaching its peak, with platforms like Maven, Reforge, and Section gaining significant traction.1 What I underestimated is how these platforms tapped into the Nike effect. Just as Nike built its brand by associating with top athletes, these platforms solved their cold-start problem by aligning themselves with top influencers in the tech and business world. This strategy not only attracted users but also established these platforms as premium learning destinations.
Seeing all this activity and interest from content creators, I assumed that there would be equal enthusiasm from learners once the hard work of creating courses was done. My confidence in the creator economy was high, largely based on what people were saying at networking events and on social media. However, I didn’t verify if there was a real market need for these services beyond the courses offered by high-profile individuals.
The Biggest Mistake
For nine months, we focused on building content and refining our production process. However, when we started getting into the sales process, we realized that quality didn’t matter. Very few customers could notice the difference in quality. Proving quality is challenging to begin with, and I suspect it wouldn’t help with sales even if we could somehow “prove our quality.” Instead of investing in making high-quality content, we should invest in distribution. It’s more effective to rely on testimonials, reviews, and social proof because people trust other people. Perceived quality is a lot more important than the actual quality.
The one thing that differentiated us was our authors’ backgrounds. They were all practitioners, and many were also consultants. We thought that turning their playbooks, built from real-life challenges, would be appealing. However, we misunderstood why people hire consultants in the first place. They want someone else to do the research, the diligence, the thinking, the hard work. When there’s a budget, companies prefer to hire someone to solve the problem directly, rather than taking the time to learn. Most of our target customers work in fast-moving industries, where today’s problem might be irrelevant in a month, or even a few days. Taking the time to learn, integrate, and apply is often not worth the struggle. It’s more efficient to hire an expert. If the consultants fail, there’s a person to shoulder the responsibility. When there’s no budget, companies prefer to do it themselves, gleaning information from blogs, podcasts, and nowadays, large language models.
Our biggest mistake was never actually testing if people wanted what we built. We simply assumed they would. I had considered starting with a simple landing page as a fake door test to gauge market interest. But this idea fell through the cracks because our team thought we needed to know exactly what we were building first. In reality, we should follow the Amazon press release strategy (write out everything we think the product will be), run small ads to validate this landing page (gather data), and de-risk our investment from there.2
If I could go back in time and change what we did, we’d first learn what our customers care about and build something entirely different. We would learn how much our customers care about networking, social proof, and competitive intelligence and build a community, conference, or networking product. Or we might decide that’s not a space we have expertise in and pivot away entirely.3
The Market Dynamic
Thanks to the Internet, everyone can discover and take the best class. The top-performing class will account for most of the platform revenue. This is what we see with Ian’s Monetization course. He’s world-class at communicating pricing concepts, breaking down case studies into everyday analogies, and engaging to listen to.
We also see this with other edtech companies. Every successful learning platform features well-known instructors who lend credibility to build the brand of the platform.4
While one of our experts has over 100k followers, we never directly sold the classes to their audience. We spent too much time planning to execute the perfect campaign and delayed promoting the book for too long. This long-term strategy also proved challenging with other authors. Some struggled to understand the need for audience testing and couldn’t pivot to create content that would appeal to this demographic. The content was set in stone, and therefore the outcome was on the wall.
Scaling a content business with multiple creators is challenging. The process of guiding each creator through strategy and execution was time-consuming. It’s more effective to minimize human input or create a marketplace model where creators can compete and thrive independently. Being an assistant for content creators wasn’t fulfilling for me.5
Beating The Dead Horse
Initially, we thought direct sales was the right approach. Our product was priced high, and we entered the market with a unique positioning, believing the market needed education.
We planned to use direct sales to get feedback on what resonated with people. Once we had our first sales, we’d scale something with proven conversion, rather than blindly running paid ads for unproven products and markets.
We never got to the point of running paid ads because we made no sales in the first few months.
We attributed the lack of sales to my poor sales skills, so we doubled down on role-playing and polishing sales scripts. We didn't question whether there was product-market fit. However, if there’s product-market fit, even terrible sales would lead to some conversion.
Meanwhile, my team kept telling us our pricing was ridiculous. We thought our course was comparable to cohort-based courses. At the time, early Maven courses were priced around $1,000, with premium ones like David Perell’s Write of Passage and Tiago Forte’s Building A Second Brain at $5,000. I insisted that our $500 price point was reasonable given the market. We even priced some courses with instructor feedback at $799. We got zero sales at this price point. After three months, we halved the price and finally saw some interest. People started putting in their email addresses and checking out the first free lecture—though still not converting.
After 100+ rejections from three months of sales efforts, I stopped focusing on what I was selling and turned my attention to what customers wanted. I shifted from direct sales to customer discovery, talking to 10 potential customers every week for a few months. When I asked about their recent learning-related purchases, it became clear they enjoyed buying books. Some collect them, others respond to effective book marketing. Regardless, there’s a strong market preference for books. This insight led us to rebrand all our courses as books and adjust pricing to $49-$69.
People like buying books
The results were immediate. Even before we officially announced the rebrand, we started receiving sales. Books seemed like the right product.
This is why Penguin Random House offers author advances and revenue shares. There’s an established process for selling books at scale, most companies have mechanisms to reimburse book purchases, and people enjoy sharing their reading habits (consider the popularity of Barack Obama and Bill Gates’ reading lists). Books also lend themselves well to word-of-mouth marketing and partnership opportunities.
The path became much clearer with this realization. I already knew the book sales playbook, so I simply followed it. However, we initially wasted time on meaningless optimization, like testing different coupons instead of giving out books for free to thought leaders to maximize word of mouth.
It wasn’t until Ellen Fishbein sent me a DM offering a free copy of her book that I realized we should do the same. If we’re truly confident about our product, a percentage of these VIPs will begin sharing and spreading the word for us. I started identifying domain experts who are active and share resources in Slack communities and began giving out our Startup Sales and Monetization book. We also began hosting events and giving all attendees copies of Fundraising and Value Proposition (which is now on Amazon). I gifted books to friends and recognized those who gave good feedback on our content creation process. This certainly helped increase our credibility. We ended up gifting 50 books in a month.
Interestingly, the fundraising book got a lot of views but was rarely shared. I hypothesize that people keep it as a secret strategy. The only ones who shared it publicly were those past the first stage of fundraising. For example, a friend offered to share our book in YC’s internal community Bookface, where most people are past the initial fundraising stage.
The Only Wise Decision
Once we demonstrated the effectiveness of books, we began partnering with established brands and communities, leveraging their audiences and credibility to sell our books. Pitching guest posts to Every worked exceptionally well. They have one of the highest-quality tech audiences interested in in-depth content, making them a perfect fit for us. Ian wrote a piece on How to Price Generative AI Products after helping many clients with this issue. The piece was instantly accepted. When it went live at 11 AM EST, I was having dinner in Zurich and vividly remember the surprise of seeing six purchases pop into my email. Sales continued for a few weeks, helping us attract our target clientele. Unfortunately, it didn’t boost other book sales, as customers were mainly there for Ian’s content. This highlighted that our cross-sell and product-led growth strategies needed improvement. We replicated this approach by making select content public, pitching to top publications, and syndicating pieces whenever possible.
For authors who excel at impromptu speaking, we hosted online workshops. I moderated at least four workshops each month, covering topics from product-market fit to fundraising. Most groups consisted of dozens of people, so it wasn’t the most scalable approach. However, each workshop built direct connections. When these experts effectively handled Q&A sessions, they instantly earned credibility.
Through our events, we discovered that people greatly value peer learning. We noticed attendees were most attentive during group introductions, likely because these invite-only events attracted equally accomplished peers. We believed this interest went beyond simple curiosity: there seemed to be significant value in connecting people at similar stages to exchange insights.
This idea explains the success of expert networks like GLG and Tegus.6 People often want to hear directly from their peers and industry insiders. During sales calls, people asked about consulting calls with experts on the Distillable platform. Although more expensive, these calls represent a familiar product that people are generally more comfortable paying for.7 We considered offering consultations to gauge how much people valued our experts but ultimately decided against competing with our authors’ existing consulting businesses.
The preference for peer learning and insider knowledge extends beyond our platform. At Google, for instance, employees prefer fireside chats or Q&A events with internal leadership, valuing immediately applicable tactics and insights over generic advice from external experts. Many perceive their challenges as unique, and managers are comfortable with letting their teams go through several learning cycles.
However, facilitating direct peer-to-peer learning in competitive industries presents challenges. Exchanging notes with colleagues doing similar work is often against corporate codes of conduct. We saw an opportunity to create a space where professionals from different companies could safely share knowledge and learn from each other.
While running these groups, I noticed that people weren’t just after knowledge. They wanted community. They sought connections with others who shared their challenges and goals. These connections inspired them to improve.
With more AI-generated content, people may trust online content less. Instead, they’ll value human connections and niche communities more. This explains why conferences and MBA programs remain popular. Their real value isn’t just the content—it’s the network of peers you build.
The AI Era
I was worried about how AI would change content creation. I kept thinking: Why would people buy books if they could just ask ChatGPT? What would AI do to writing?
These questions could have stopped me in my tracks. But instead, I saw a chance to do something new. Our content business already felt old compared to what AI could do. So rather than trying to fix our old way of doing things, I started playing around with AI editing.
We began by using AI to do the boring stuff in creative work. Things like proofreading, writing subject lines, and copy editing. But we still spent most of our time on the big-picture editing. That’s what helps experts share insights they might not even know they have. We found the best way to do this was to interview the author, let them talk freely, and record it all. This gave the AI enough info to learn how the author talks and writes. Sometimes just five minutes of human speech was enough.
I tested AI writing a lot, both for myself and for influencers. I liked it because the worst part of my job was dealing with people’s schedules. Working with humans is great for creating new ideas together. But people get sick, go on vacation, change their minds, or get difficult to work with. This can make projects take way longer than they should. I learned to be clearer about what I need and to point out problems early.
At first, my team said my AI-written essays sounded robotic and boring. But when I started recording myself talking and only used AI for editing, things got much better. I stopped trying to make everything perfect. I realized that content that’s too polished can seem fake. Instead, I focused on sounding like a real person (myself!). We used this same idea for writing social media posts too.
I also started helping businesses use AI. I helped Din Tai Fung use AI in their company, and I began working with hospitals too. I quickly saw that this could be something big, not just a side project. So I decided to end our content business.
It was a big change, and I wasn’t sure what would happen. But it felt right. AI was opening up a whole new world, and this time, I wanted to create change, not just react to it.
This whole experience taught me that success in startups isn’t just about having a great idea or building a cool product. It’s about really understanding what people want, being ready to change when things aren’t working, and creating value by connecting people. As I finished up with Distillable and looked to what was next, I knew these lessons would help me in whatever I did next.
Maven, for instance, started their platform with well-known figures like Sahil Lavingia, Li Jin, and Lenny Rachitsky. Reforge began by offering in-person workshops taught by Andrew Chen in San Francisco. Section built its brand around Scott Galloway and his network.
We could have redirected the production budget to run ads on a fake landing page with a waitlist. If we couldn’t get 50 pre-registrations or deposits, we should have reconsidered whether to develop the courses at all.
Thankfully, my team got involved in questioning the decisions I made. They pushed me to think through the goal of each action and not waste time on things that have no value. Their line of questioning was rational, honest, and truthful.
As mentioned earlier, most edtech companies leveraged influencers to bootstrap their early growth. Masterclass, for example, built their entire brand around celebrities.
Instead of continuing down this path, I pivoted to build my expertise in productivity. I picked up coding and writing again. In a few months, I received a LinkedIn Top Voice badge and won several hackathon awards. More importantly, it paved the way for my next venture, which I’ll cover in the upcoming chapters.
GLG and Tegus are expert booking platforms that connect you with industry insiders. A founder once compared us to Tegus, which transcribes consulting calls and incorporates the expert library into their product. I had considered making the course creation process a product itself, as many raw, insightful conversations occurred during this stage. However, I didn’t pursue this because it didn’t fit our main product. Also, too many hot takes to be edited out.
For example, Intro is a marketplace that allows anyone to book experts at a premium price. One of their top experts, Nikita Bier, charges $1800 for a 15-minute call.