Lecture 4: Building Product, Talking to Users, and Growing

What this lecture covers

Adora Cheung, co-founder of Homejoy, gives a practical, mistake-driven walkthrough of how to take a startup from zero users to steady growth. Drawing on the many pivots Homejoy went through before landing on its cleaning-services idea, she covers validating a problem, immersing yourself in an industry, building a minimum viable product, collecting honest feedback, and choosing a growth strategy once a product is ready to scale.

After watching, you should be able to describe how to manually recruit early users, what makes user feedback trustworthy versus flattering, and how to weigh sticky growth, viral growth, and paid growth against each other using retention curves and customer acquisition cost.

Key ideas

• Immersion over scattered effort: block out concentrated stretches of time on a startup rather than a couple of hours a day, since context switching wastes the deep focus a startup idea needs.
• Validate the problem before building: state the problem in one sentence, confirm you are personally passionate about it, and verify other people have it too, before writing code.
• Become an expert in the industry: Cheung and her co-founder trained as cleaners themselves and researched competitors exhaustively before building Homejoy, which revealed inefficiencies they could exploit.
• The honesty curve: feedback from people close to you (family, friends) is unreliable because they want to support you; feedback from strangers, especially paying strangers, is the most honest.
• Minimum viable product means viable: build the smallest feature set that actually solves the core problem, not a stripped-down product that fails to work at all.
• One clear positioning statement: a simple, benefit-focused sentence like "get your place clean for $20 an hour" converts far better than a long feature list.
• Recruit users manually at first: go to online and local communities, or physically show up where potential users are, rather than paying for ads before you have product-market fit.
• Three types of growth: sticky growth (existing users return and spend more), viral growth (users refer others), and paid growth (buying users, sustainable only when customer lifetime value exceeds acquisition cost).

Walkthrough

Immersing yourself and avoiding the novice cycle (0:01)

Cheung introduces the "novice approach": building in secret, launching on TechCrunch, getting a burst of users who don't stick around, and eventually giving up. She recommends concentrated blocks of focused time on a startup idea instead of scattered effort, and stresses validating the problem statement and your own passion for it before building anything, illustrating this with Homejoy's own year-long detour building a platform for life coaches and therapists that neither founder actually wanted to use.

Becoming an expert in the industry (5:07)

Once you have a validated problem, Cheung recommends immersing yourself in the target industry, even taking a job in it, as she and her co-founder did by training and working at a cleaning company before building Homejoy. This immersion revealed why local cleaning companies stayed small: scheduling and booking were done inefficiently. She also describes exhaustively researching competitors, down to reading their financial filings, to reach a level of expertise that earns customer trust.

Defining customers and storyboarding the experience (9:13)

Cheung explains the value of narrowing to a specific customer segment early rather than trying to serve everyone at once, and recommends storyboarding the entire user journey, from how someone discovers the product to what they experience after using it, before writing any code.

Building the MVP and getting first users (11:17)

She stresses that "viable" in MVP is often skipped, leading to products that don't actually solve the problem. A clear, simple positioning statement, illustrated by Homejoy's shift to "get your place clean for $20 an hour," made a major difference in conversion. Early users should start with people you already know, then expand to online and local communities; Cheung recounts personally recruiting strangers at a Mountain View street fair by handing out cold water bottles on a hot day.

Collecting honest feedback (17:20)

Cheung recommends talking to users directly rather than relying only on surveys, treating these conversations as relaxed rather than interrogations. She introduces the "honesty curve": feedback from people close to you is unreliable because they want to support you, while feedback from strangers, especially those who paid, is far more trustworthy. She also describes tracking retention and using net promoter score as a faster leading indicator than long-term retention data.

Building for the next stage of growth (23:26)

Cheung advises optimizing for the next stage of growth rather than a hypothetical future scale, favoring manual processes over premature automation, since manual work reveals what should eventually be automated. She introduces "temporary brokenness is better than permanent paralysis," meaning founders should not obsess over edge cases early on, and warns against the "Frankenstein approach" of building every feature a user requests without first understanding the underlying problem behind the request.

Choosing a growth strategy (29:32)

Cheung describes three types of growth: sticky growth, measured through cohort retention curves and customer lifetime value; viral growth, which depends on strong product experience combined with well-placed referral mechanics; and paid growth, which is sustainable only when customer lifetime value exceeds customer acquisition cost. She walks through a cohort analysis example and a CAC calculation, and stresses keeping payback time short, around three months for safety, since overspending relative to future revenue can bankrupt a young company.

The art of pivoting (45:55)

Cheung explains that Homejoy was the thirteenth idea she and her co-founder tried before it worked. She offers a growth-plan test: if you are executing hard and see three or four weeks of flat or declining growth, that's usually a sign to reconsider the idea, distinct from a temporary lull that recovers with continued effort.

Before you watch

• Watching Lectures 1 through 3 first gives useful context on idea validation and founder mindset that this lecture builds on with hands-on execution advice.
• Basic familiarity with terms like MVP, cohort analysis, and customer acquisition cost helps, though Cheung defines each as she goes.

Check your understanding

1. Why does Cheung consider feedback from close friends and family less reliable than feedback from paying strangers?
2. What is the "novice approach" to launching a startup, and why does it tend to fail?
3. How does Cheung use CLV and CAC to decide whether a paid growth channel is worth pursuing?
4. What signal does Cheung use to decide whether a slow-growing idea needs a pivot rather than more patience?