In my years launching products like Starling Bank, Skype for mobile, and Yolt, I've discovered that finding true product-market fit rarely comes from perfect systems or automation - it comes from being willing to do the messy, manual work that simply won't scale. 

‍

Time and again founders often get trapped optimising too early, wanting million-user infrastructure before finding their first hundred true fans. 

‍

The reality? The path to the right audience, the right channels, and the right growth levers almost always runs through unscalable, hands-on customer engagement that no algorithm can match.

This week, I'm diving into why this counterintuitive approach remains the most reliable way to build something people actually want.

‍

The Paradox of Early-Stage Growth

Don't be afraid to do things that don't scale.

‍

This classic advice from Y Combinator co-founder Paul Graham feels counterintuitive in today's efficiency-obsessed business culture, but it remains one of the most valuable principles for startups and growing businesses.

‍

Time and again, we speak with founders fixated on building "scalable processes" from day one. Their common refrain: "But it won't work at scale..." This becomes the perfect excuse to avoid the messy, manual work that often drives early momentum.

‍

This mindset is a dangerous trap. Some of the world's most successful companies began by doing things that would make efficiency experts cringe:

‍

• Stripe: The founders, Patrick and John Collison, would personally visit customers' offices to assist with integration, sometimes actually writing the code for them.
  
• Deliveroo: Will Shu, the founder, spent Deliveroo's first year as one of the primary delivery riders, personally delivering orders while simultaneously running the company.
  
• Airbnb: The founders went door-to-door in New York, helping hosts take professional photographs of their properties and offering personalized advice on pricing and descriptions.
  
• Gymshark: After a website crash during Black Friday in 2015, founder Ben Francis personally wrote 2,500 handwritten apology notes with discount codes to affected customers.
  
• Zappos: Tony Hsieh initially created a website with images of shoes from Foot Locker. When someone ordered, he would buy the shoes at retail price and ship them himself—a completely unsustainable business model that nevertheless allowed him to test his concept.
  
• DoorDash: The founders started by printing menus from local restaurants, creating a simple website, and delivering the food themselves using their personal phones as the dispatch system.

‍

Why Unscalable Actions Create Disproportionate Results

As Graham explains in his essay, doing things that don't scale is essential because:

‍

1. It creates an exceptional user experience. When you manually tend to each customer's needs, you create advocates, not just users. The personal touch makes people feel valued in a way that automated systems rarely can.
   
2. It provides unfiltered market feedback. Direct interaction with customers reveals insights that surveys and analytics can't capture—the emotional reactions, unexpected use cases, and unspoken pain points that shape product development.
   
3. It helps you discover your ideal customer profile. By serving customers manually, you learn exactly who benefits most from your solution, allowing you to refine your targeting.
   
   

The Deliberate Constraint of Starting Small

One counterintuitive "unscalable" approach that pays dividends is deliberately constraining your market. As Graham notes in his essay, Facebook exemplifies this strategy:

‍

"Facebook started out exclusively for Harvard students, who thought it was cool that it was exclusive... After Facebook became popular at Harvard, they opened it up first to other Ivy League schools, then to all colleges, then to high schools, and eventually to everyone. But by that time they had the critical mass of users and the site was good enough that the first round of users stuck around."

‍

Stripe followed a similar path by initially focusing exclusively on developers at YC startups. Notion began by targeting design-forward technology companies before expanding to broader business use.

‍

It's like controlling a fire: concentrate your energy on a small area until it burns hot enough to sustain itself when you add more fuel.

‍

The Dangers of Premature Scaling

In our eagerness to grow efficiently, we often make two critical mistakes:

‍

1. Building things people don't want

You don't need a fully formed product to launch. Waiting too long means you risk spending resources developing the wrong solution.

‍

Consider Ally Fekaiki of Juno, who started with just a landing page selling his vision for employee benefits. This minimal approach allowed him to validate market interest and gather critical feedback before building his full platform.

‍

The key: Start small and gather data on what excites people. Build once you know what they want.

‍

2. Automating processes that don't work

Imagine an onboarding email sequence that's beautifully automated but fundamentally ineffective. Automation simply scales your mistakes faster.

‍

Instead, take the hands-on approach initially:

‍

• Personally email new customers
• Call users who seem stuck
• Walk people through onboarding
• Have actual conversations about their experience

‍

Only once you've discovered what resonates should you automate the process.

‍

Recent Examples of Unscalable Success (2023-2024)

Superhuman

Rahul Vohra's email app Superhuman charges $30/month and until recently required users to go through a personal onboarding call with a team member—despite having over 100,000 users on their waitlist. This "high-touch" approach ensured users understood the product's value and became immediate advocates.

‍

Replicate

This AI research company personally reviewed early users' projects, offered direct implementation support, and even paired engineers with customers for one-on-one consultations—practices that helped them build exceptional loyalty in the fiercely competitive AI infrastructure space.

‍

Maven

This cohort-based learning platform started with founder Gagan Biyani personally recruiting instructors and helping design their courses. Rather than building elaborate course creation tools first, they focused on ensuring exceptional course quality through manual intervention.

‍

Figma's Customer Success

Even as they grew to a multi-billion dollar valuation, Figma maintained a practice where product managers and engineers would personally join customer calls to understand pain points—a practice that contributed to their exceptional product-market fit.

‍

The B2B Application: Quality Over Quantity

The same principles apply powerfully to B2B sales. Before thinking about your 1,000th customer, obsess over delighting your first 10:

‍

• When a new lead signs up, call them personally
• Understand their specific use case and customize your pitch
• Discover the unexpected insights that only emerge in conversation
• Manually implement solutions to their problems

‍

As investor Jason Lemkin notes: "Your first 10 customers will define your company." These early relationships set the tone for your product development, marketing messaging, and company culture.

‍

The Path Forward: Excellence Before Efficiency

As Paul Graham says, "It's not enough just to do something extraordinary initially. You have to make an extraordinary effort initially."

‍

The true formula is:

‍

1. Do things that don't scale to create exceptional experiences
2. Learn what works through direct customer interaction
3. Perfect your approach based on real feedback
4. Only then systematize and automate

‍

The lesson is clear: Don't let the fear of scaling prevent you from taking bold, unscalable actions that drive growth. The path to success often begins with doing the hard, manual work that others aren't willing to do.

‍

In a world obsessed with automation and efficiency, your willingness to do things that don't scale might be your greatest competitive advantage.