Adora Cheung on Homejoy, YC, Vote-by-Mail, and Instalab

概览

This episode follows Adora Cheung from the earliest PathJoy idea through the rise and failure of Homejoy, her years as a YC partner, an under-the-radar 2020 vote-by-mail project, and her current company, Instalab.

A central thread is Adora’s willingness to learn from the ground level: cleaning houses herself to understand Homejoy’s supply side, analyzing voter behavior like a conversion funnel, and using her own health problems as the starting point for Instalab.

The biggest startup lesson is that growth can hide a broken product only temporarily. Homejoy scaled across many cities before retention and operations were strong enough, then worsened the problem through price competition. Adora frames the failure as a set of avoidable decisions rather than a single fatal event.

The episode also highlights a consistent operating style: get close to users, track the full funnel, avoid vanity metrics, and build systems that make important behavior easier.

分段落总结

[00:00] Introducing Adora Cheung

[事实] Jessica Livingston and Carolyn Levy introduce the podcast as conversations with successful Silicon Valley founders about how they built their companies. [事实] Adora Cheung is introduced as a YC founder from 2010, the founder of PathJoy/Homejoy, a former YC partner for five years, and now the founder of a new company. [推测] The episode is framed as both a founder story and a retrospective on hard-earned operating lessons.

[01:07] From PathJoy to Homejoy

[事实] Adora says PathJoy began around 2009 as a marketplace for services, initially focused on therapists and life coaches. [事实] The company attracted supply but could not generate enough demand, so the team pivoted repeatedly during and after YC. [事实] After roughly two years and about a dozen pivots, the company became Homejoy, a marketplace for home service providers, best known for cleaning. [推测] The PathJoy experience taught the founders to validate demand before overbuilding the supply side.

[02:29] Finding the Cleaning Problem

[事实] Aaron, Adora’s brother and co-founder, struggled to find a good cleaner, which led the team to test whether others had the same problem. [事实] Before Homejoy, the team made small amounts of money through content and SEO-oriented websites, including an entertainment news site. [事实] The first Homejoy customers came through a simple website, phone number, posters, and local outreach in Mountain View. [事实] Adora cleaned the first large customer house herself because they had not yet found a cleaner; the job took more than 12 hours. [推测] The first cleaning job exposed the gap between “cleaning” as a concept and the operational skill required to deliver it professionally.

[06:03] Learning to Clean by Doing

[事实] Adora tried to get cleaning jobs so she could learn the work directly, and most companies rejected her because she had no cleaning experience and seemed overqualified. [事实] She eventually got a cleaning job in San Francisco while living in Mountain View, sometimes arriving very early and sleeping in her car before the workday. [事实] She learned that efficient cleaning depends on the right tools, the right cleaning chemicals, sequencing work from top to bottom and inside to outside, and splitting work well with a partner. [事实] She worked as a cleaner for about a month. [推测] This period helped her understand both how to deliver the service and how to evaluate and onboard cleaners for Homejoy.

[09:41] Bootstrapping the Marketplace

[事实] Homejoy recruited cleaners through Craigslist and in-person conversations, while trying to convince them that the company had customers and would pay reliably. [事实] On the demand side, the company used a website, a visible phone number, flyers, and word of mouth before raising a proper seed round. [事实] Adora says having a phone number mattered because customers could hear a real person from an unknown company. [事实] Aaron was especially strong at customer acquisition, ad funnels, and lowering customer acquisition costs. [推测] The early marketplace worked because the team built trust manually before trying to scale acquisition.

[12:23] Family, Startups, and Siblings

[事实] Adora says she did not always know she wanted to start a company with Aaron. [事实] Aaron became interested in startups at MIT, while Adora discovered startups later in graduate school after studying computer science as an undergraduate and economics in graduate school. [事实] Adora grew up in South Carolina and says Silicon Valley was not part of her world at the time. [事实] Her mother, from an immigrant family background, was not happy that Aaron chose a startup path instead of a traditional prestigious job. [事实] Adora’s current co-founder is another brother, Alex. [推测] The family stories show how unconventional the startup path felt compared with the expectations around education and stable careers.

[15:53] Fundraising and Being Taken Seriously

[事实] Paul Graham once tweeted Homejoy’s revenue growth chart anonymously, after which investor interest increased. [事实] Adora says she does not know exactly why investors were initially uninterested, but she does not think the cleaning business itself was the problem because on-demand models were popular. [事实] She recalls that investors often directed questions to Aaron, even though she was the CEO and could answer technical, product, and business questions. [事实] She eventually asked Aaron to stop coming to some meetings so investors would focus on her as the company’s leader. [推测] Gender bias likely played some role in how investors engaged with the team, though Adora does not claim certainty.

[17:24] Why Homejoy Failed

[事实] Adora says Homejoy “rapidly scaled a broken product.” [事实] The product worked well in the early San Francisco and Los Angeles markets, but scaling required more features and stronger operational systems. [事实] Retention started strong and then declined; Adora says founders should stop and fix the product when retention falls below an important threshold. [事实] Homejoy grew to almost 300 employees, not including contractors, and expanded into many cities, including in Europe. [推测] Homejoy’s growth created organizational complexity faster than the product and operations could absorb it.

[20:06] Price Wars and Bad Growth

[事实] Homejoy entered a price war with competitors and offered $19 cleanings. [事实] Adora says the company could not make money on $19 cleanings because cleaners still had to be paid. [事实] The theory was that discounted first cleanings would lead to repeat bookings, but poor retention meant many follow-up cleanings did not happen. [事实] Adora says startups should know what competitors are doing but should not measure themselves against competitors instead of building a high-quality product. [推测] The price war amplified the retention problem by buying customers before the core experience was strong enough.

[22:17] Recognizing the End

[事实] Adora says there were disagreements inside Homejoy about whether to stop growth and fix the product. [事实] Some people raised alarms earlier, but she says she did not make the decision in time. [事实] By the end, she and Aaron were largely aligned on the outcome, lessons learned, and what they would do differently. [推测] The failure gave Adora a clearer framework for advising later founders at YC.

[23:03] Lessons from Being a YC Partner

[事实] Adora says giving advice is easier than taking it and acting on it. [事实] From advising startups, she learned to recognize founder qualities such as communication, perseverance, and the ability to keep moving through problems. [事实] She cites the “be a goldfish” idea: good founders absorb the lesson from bad events and move on quickly. [事实] She says founders who dwell too long on setbacks can slow their decision-making and stall the company. [推测] Adora’s YC advice centers on decisiveness, emotional recovery, and practical forward motion.

[25:07] The Startup School Email Mistake

[事实] Adora enjoyed reading YC applications and worked on Startup School when it had an application process for its mentoring component. [事实] Around 2017, during Demo Day, acceptance and rejection emails were accidentally reversed for Startup School applicants. [事实] Jared suggested letting everyone in, and Adora eventually agreed after considering Startup School’s mission to teach entrepreneurship broadly. [事实] That mistake helped lead to Startup School becoming open to anyone who signed up. [推测] A serious operational error became a product and mission shift that expanded access.

[28:37] YC Batch Memories

[事实] Adora’s YC Summer 2010 batch included people she remained close with, including Yuri Sagalov, Adam Goldstein, Solomon, and Francis and Ed from Tapzilla. [事实] After the batch, some founders recreated YC-style dinners and social gatherings around Mountain View and Palo Alto. [事实] Adora remembers Demo Day in the old Pioneer Way office as crowded, hot, and split into multiple sessions. [事实] During one pitch, she was distracted after noticing Demi Moore in the audience. [推测] The memories contrast early YC’s scrappy physical constraints with the more organized Demo Day format the hosts describe later.

[32:08] Applying Startup Skills to the 2020 Election

[事实] After Trump’s 2016 election, Adora began thinking about how to use her skills to help prevent his reelection and support Biden. [事实] YC allowed her to take time off to work on the problem. [事实] She first went to Wisconsin, where her parents lived, and talked to voters through Craigslist postings and coffee meetings. [事实] She realized direct voter conversations would not scale and later focused on vote-by-mail during COVID. [推测] Adora approached political turnout like customer discovery before moving toward a scalable funnel.

[35:30] Building a Vote-by-Mail Funnel

[事实] Adora found that vote-by-mail registration websites were often clunky and hard to use. [事实] She cites Florida as an example where form issues could block people, including problems with accented names. [事实] Election offices provided data showing whether people registered for vote by mail, whether ballots were sent, and whether ballots were returned. [事实] Adora worked with Tech for Campaigns, recruited Mark Lindsay to run much of the tech work, and helped optimize ads and the full funnel. [事实] The team tested in the Florida primaries, then focused on Wisconsin, Pennsylvania, Minnesota, Michigan, and Florida for the general election. [推测] The project treated voting as an end-to-end conversion problem rather than stopping at ad clicks.

[40:13] Impact and Nonpartisan Mechanics

[事实] The hosts point out that easier vote-by-mail access is structurally nonpartisan because it helps people vote, regardless of party. [事实] Adora says the best-performing ads were nonpartisan reminders about deadlines. [事实] The team tracked and followed about 270,000 voters through to submitting an actual vote. [事实] Adora says they got about 50,000 votes in Pennsylvania, where the winning margin was around 80,000, and about 40,000 in Wisconsin, where the winning margin was around 20,000. [事实] She says they cannot know how many votes they were fully responsible for and tried to avoid cannibalizing other initiatives. [推测] The numbers suggest the project may have mattered most in close swing states, but the transcript does not establish exact causal impact.

[41:51] Metrics, Donors, and Government Forms

[事实] Adora’s takeaways are that voter turnout is complex but does not need to be extremely hard or costly with the right mindset and data. [事实] She says people often do want to vote, and the job is to make the process easier from start to finish. [事实] She argues big donors should avoid giving money to organizations that cannot show metrics-driven efficiency. [事实] She says voter registration and vote-by-mail government forms need major improvement. [事实] She tested meme-style ads from another organization and found they had high click-through rates but almost no follow-through to registration. [推测] Adora’s critique is that political spending often optimizes for visible activity instead of final outcomes.

[44:03] Introducing Instalab

[事实] Instalab is an at-home blood testing service that coaches people on optimizing their health and reaching peak health. [事实] Its target customers are busy people, including founders, executives, and working parents. [事实] Adora started the company because her own health declined after Homejoy. [事实] She says busy founders often know they should focus on health but make the company the top priority. [推测] Instalab applies Adora’s marketplace and funnel thinking to preventive health behavior.

[45:01] Health Debt After Startup Life

[事实] After Homejoy, Adora had gained weight and had high lipids, high cholesterol, hypertension, and metabolic issues. [事实] She says her diet during startup life was based on convenience and energy, including too much caffeine and four or five Red Bulls a day. [事实] She describes “life debt” by analogy to technical debt: accumulating problems early can make future health issues harder to fix. [事实] She believes she could have addressed issues earlier if she had known they existed and how to fit changes into her lifestyle. [推测] The company’s mission is partly a response to the hidden costs of founder intensity.

[47:34] How Instalab Works

[事实] Instalab sends someone to a home or office for a blood draw and measurements such as blood pressure, weight, and grip strength. [事实] Adora says the visit takes about 15 minutes and sometimes less. [事实] The company provides results and explains where the patient is, then suggests a few next steps. [事实] Its panel includes 60 biomarkers, and common issues include metabolic problems, cardiovascular issues, and high blood pressure. [事实] Adora says close to 100% of patients find something they did not know was an issue. [事实] Instalab operates in New York, San Francisco, Los Angeles, and Austin. [推测] The service is designed to reduce friction by bringing testing to people instead of asking busy patients to seek it out.

[51:52] Feedback Loops for Health

[事实] After results, patients can talk to a longevity physician if they have more questions. [事实] Instalab recommends retesting in about three months if patients act on recommendations. [事实] Adora says three months is enough time to see many biomarkers move in the right direction. [事实] She contrasts this with traditional healthcare, where people may get a basic lipid panel and then wait years to measure progress again. [推测] The founders’ attraction to data and feedback loops is part of why Adora believes the product fits startup people.

[53:15] Hiring Phlebotomists and Learning from Homejoy

[事实] Instalab hires full-time W-2 employees instead of 1099 contractors. [事实] Adora says she is done with the 1099 model because it limits the company’s ability to control quality and standardize how work is done. [事实] Instalab’s higher price point allows it to absorb the costs of full-time employment. [事实] Reliability remains an important hiring criterion, just as it was at Homejoy. [事实] Adora personally tests phlebotomists by having them draw her blood because she wants the experience to be as painless as possible. [推测] Instalab repeats some of Homejoy’s in-home operational complexity but tries to avoid one of Homejoy’s quality-control weaknesses.

[55:26] Baby Steps and Behavior Change

[事实] Adora says many Instalab patients are vitamin D deficient. [事实] One suggested “baby step” is to turn an existing phone call or meeting into an outdoor walking call to get sunlight. [事实] Carolyn compares this incremental approach to Noom’s food-change strategy. [事实] Adora mentions Supportive, an app where users can share scale data with friends for encouragement. [推测] The health strategy is not framed as dramatic life redesign, but as small edits that fit into a busy schedule.

[58:33] Closing Reflections

[事实] The hosts say they enjoyed catching up with Adora and learning more about Instalab. [事实] Jessica says she rarely asks founders about failure because many people are not honest about why they failed, but she trusts Adora’s self-assessment. [事实] Carolyn says extracting lessons from failure is important for being a YC partner. [事实] The hosts describe Adora as a founder at heart who wants to build, make, change, and improve things. [推测] The episode presents Adora’s career as a through-line of direct learning, operational rigor, and willingness to revisit painful lessons.

播客点评/总结

This episode is valuable because it goes beyond a simple success story. Adora gives a candid account of Homejoy’s failure, especially the danger of scaling before retention and operations are truly working. The discussion is useful for founders because it connects abstract startup advice to concrete decisions: city expansion, discounts, retention thresholds, hiring models, and funnel measurement.

The strongest parts are the stories where Adora gets close to the work: cleaning houses herself, talking to voters in Wisconsin, testing vote-by-mail ads against actual ballot return data, and using her own health metrics to shape Instalab. These examples make the episode feel practical rather than theoretical.

[推测] A limitation is that several important topics are discussed from Adora’s perspective without outside verification, especially the exact political impact of the vote-by-mail project and the current performance of Instalab. The transcript also does not deeply cover Homejoy’s legal or labor issues beyond the contractor-versus-employee distinction raised later.

[推测] This episode is best for startup founders, operators, YC followers, and people interested in marketplace execution, growth mistakes, political technology, or founder health. It is especially relevant for listeners who want honest postmortems rather than polished founder mythology.