Homejoy at the Unicorn Glue Factory
Will the home-cleaning revolution be Uberfied? How one company tried, and spectacularly failed.
It was Thanksgiving 2013 and the San Francisco offices of Homejoy were uncommonly quiet.
The online home-cleaning startup was in the midst of an explosive expansion and only shut its doors two days each year: Thanksgiving and Christmas Day.
But on this particular holiday, a booking had slipped through unnoticed, due to a website malfunction, according to a former employee. Rather than cancel an appointment at the last minute, Homejoy’s cofounder and CEO Adora Cheung grabbed a toilet brush and a vacuum cleaner. Then she headed to San Francisco’s Mission Dolores neighborhood and scrubbed.
“The customer had no idea,” said Arjun Naskar, one of the earliest employees at Homejoy, who shared the story as an example of the founders’ intense determination to succeed.
Home cleaning represents an estimated $400 billion global market that many believe is ripe for an Uber-like overhaul. Homejoy was one of the first of its kind to market. Using logistics algorithms to easily connect homeowners with contract-for-hire cleaners and schedule visits, it drew the attention of venture capitalists hunting for new “gig economy” unicorns and eventually raised some $40 million from Google Ventures and PayPal cofounder Max Levchin, among others.
Less than two years later, it hit the wall. After a period of torrid growth, the founders suddenly announced they had run out of money and had to shut the doors for good. Months of triage had produced no new investor or potential buyer, and seemingly overnight the one-time darling of Sand Hill Road became, not just another failed high-risk venture, but a stark cautionary tale for the entire on-demand economy.
Trouble With the Groupon Customer
When Homejoy folded, a slew of media articles pointed to worker classification lawsuits that plagued the company in its final months.
Like Uber, TaskRabbit and other well-known on-demand economy companies, Homejoy treated its cleaners as independent contractors, and not employees, despite how many hours they worked. Some litigators did not agree with this assessment, arguing that Homejoy and its ilk were depriving workers of reimbursements and overtime wages. At the time the company shut down, it was facing four employment suits challenging its workers’ status, and a judge had just handed class action status to a raft of suits brought against Uber by its drivers. The contract-for-hire system — key to the cost structure and profits of the on-demand model as currently conceived — was suddenly teetering.
But was that the reason for Homejoy’s collapse? At the time, Cheung told the technology blog Recode that the lawsuits were the “deciding factor” in Homejoy’s failure to raise additional funding. Others paint a different picture. In interviews with more than a half-dozen former employees who spoke with Backchannel for this story, a more complicated story emerges. The lawsuits were not the primary nor the proximate reason for the company’s demise, these people assert.
In fact, Homejoy was grappling with far more immediate problems that might have deterred potential investors equally or more: mounting losses, poor customer retention, a costly international expansion, run-of-the-mill execution problems, technical glitches and the steady leak of its best workers to direct employment arrangements with its own (now former) clients.
One of its biggest problems was the crippling cost of customer acquisition. By mid-2014, thousands of people were scooping up deeply discounted first time Homejoy cleanings for $19.99 on daily deal sites like Groupon. The company offered these aggressively even though its own internal data showed most of these people never used the service again, according to three ex-employees.
Former West Coast operations manager Anton Zietsman said that Homejoy was all-too aware of the challenges for startups and small businesses to attract repeat customers from Groupon. But he said they were forced to rely on it heavily because of intense competition with their chief rival, Handy, which employed a similar strategy.
A third-party analysis of the company’s financials viewed by Backchannel showed that only about a quarter of its customers continued to use the service after the first month, and less than 10 percent used it after six months. (The source of this report requested anonymity because the data is proprietary and not authorized for public release.)
“The key problem is that we weren’t making enough money on our customers,” recalled Daniel Hung, the second full-time engineer to join the company. “We were spending a lot of money to acquire them, but not really retaining them.”
At the same time it was racking up steep losses to churn, in the first months of 2014 Homejoy cut its standard prices drastically to attract customers in dozens of new cities and drive growth.
Cost was not the only problem. According to Zietsman, Homejoy also struggled to create a reliable service. Many first-time customers were not satisfied with the cleaning, or experienced a last minute cancellation. “We didn’t figure out how to deliver a consistently high-quality service,” he said.
By the time the founders began to explore strategies for improving service and alternatives to Facebook ads and Groupon bargains to yield higher retention rates, it was already too late.
At the end of 2014, in-house marketing attempted to right the ugly retention curve by prioritizing paid search and email campaigns. The idea was to attract customers who were already looking for a house cleaner, rather than bargain-hunters, said Hasan Luongo, a former growth marketer for the company who now works as a director of growth at YourMechanic. The client services teams also started experimenting with strategies to reward the best cleaners. Despite these efforts, the service shut down just six months later.
“Each individual home cleaning was simply not profitable,” said Jeremiah Owyang, the founder of Crowd Companies, a research firm that focuses on the on-demand economy. “Homejoy was essentially operating at a net-negative.”
Cheung did not respond to requests for an interview for this story. Investors Google Ventures declined to comment, and Max Levchin did not respond to inquiries.
The Obsession With Growth
Homejoy is far from being the lone example of a Silicon Valley startup to prioritize growth over revenue. Silicon Valley’s investors have published countless blog posts urging entrepreneurs to focus on expanding the customer base and to worry about a sustainable business model later.
“Growth is where all the action is, and to where all the money flows,” wrote Jeff Jordan from Andreessen Horowitz, one of the venture capital firms that invested in Homejoy at the seed stage.
Many on-demand economy companies rely on these methods. Uber, the largest and most successful of the group, has spent aggressively on marketing and ride subsidies to woo customers. Academics and experts say it could take years before Uber’s business model is viable on a global scale.
In retrospect, expanding too quickly into new markets proved to be a major challenge for Homejoy, and put the entire company at risk.
At its height of popularity, Homejoy was operating in more than 30 cities, including Los Angeles, London, Berlin, and New York. Naskar, the early employee who was responsible for growth, lived a peripatetic existence. Whenever he entered a new city, Naskar said the marketing team would crank out bargains to attract first time customers, who would flood the site.
In the spring of 2013, the company had about 20 employees. Two years later, it had hired more than 100 people across the world, including city managers, customer support, and an enviable engineering team poached from giants like Facebook and Google. “When I joined in 2014, there were a lot of new faces,” said Zietsman. “We were growing at a steady clip.”
The company had come a long way since siblings Adora and Aaron Cheung founded “the Uber for cleaning” in 2010, attracting a small amount of money from the competitive startup accelerator program Y Combinator. Many of the first employees, including Naskar and Hung, were friends of Aaron from MIT. But the startup would later attract top executive talent, including a former vice president from Starbucks and a short-lived CFO from Zynga, Mark Vranesh. Sources say Vranesh left after it became clear that the company would not be taken public.
The growth was initially exciting to investors, but it didn’t take long for cracks to begin to show. It became apparent that some cities were not ideal for a service like Homejoy. According to Naskar, in less dense cities, like Tampa Bay, Florida or Atlanta, Georgia, it was difficult to recruit sufficient numbers of customers. And in some of these regions, the competition for cleaners from well-established professional cleaning services, like MerryMaids, was fierce.
“Competitors were undercutting each other with cheaper and cheaper deals,” he said. “The economics weren’t all that favorable.”
It remains to be seen if Homejoy’s rivals will fare better in the long run. Handy declined to comment. Although both companies employed many of the same tactics, such as offering cheap deals on Groupon, Handy’s executive team made the strategic decision to expand to fewer cities and operate with a slower burn rate than Homejoy. Moreover, Handy attracted more than $60 million in funding to Homejoy’s $40 million, which gives it a longer runway. A MerryMaids spokeswoman also declined to comment, citing a “quiet period.”
In particular, Homejoy’s aggressive international push may have hastened its demise.
Jennifer Miksch, who worked in operations and product management, started working for Homejoy in London before relocating to Berlin, where she said it was difficult to keep up with the level of demand for cleaners. In London, she faced the opposite problem. “American best practices were useless,” she said.
Moreover, by focusing on growth above all else, other projects to boost retention and reduce costs fell to the wayside. Zietsman said management was spread too thin trying to span multiple languages and countless geographies, and could not focus on basic supply chain and operations.
For example, according to Zietsman, Homejoy city managers were tasked with buying up and distributing cleaning supplies. But the company did not get around to determining which products were effective and low cost, or in forging deals with suppliers. It also failed to keep track of cleaners in order to retrieve unused supplies Homejoy had purchased on their behalf. “It was a money pit,” said Zietsman.
Another former employee, who requested anonymity, recalled a persistent issue that took a backseat to growth. For months, the person said, the founders failed to resolve a flaw in the algorithm, which set up back-to-back jobs for cleaners without accounting for the transit time. Cleaners traveling from Brooklyn to New Jersey would often be allotted 30 minutes to cross Manhattan, for instance, which New Yorkers know is a near-impossible feat. Despite repeated requests from the client services team, which handled complaints from the cleaners, it took months before the engineering team prioritized updating the algorithm.
In the meantime, stressed-out cleaning professionals would show up late to appointments to face the wrath of customers. An already-strained customer services team needed to dole out countless discounts in the hopes of placating customers.
“We were bleeding cash across the globe,” said Zeitsman. “All to focus on the growth we thought we needed to justify our valuation.”
Talent and Culture
If execution was often a problem, effort was not. It wasn’t uncommon for a Homejoy employee to pull a 14-hour shift at the San Francisco office. At 11p.m, Luongo said there were often 20 people still at their desks.
The company prided itself on employees who were willing to pull their weight and get their hands dirty — and that directive came from the top. One initiation for new hires, top executives and engineers alike was to have them go along on a “test clean.” Stephanie Toler, a former operations manager who previously ran a small cleaning business, helped train some new hires and developed a playbook of sorts for employees to learn how to properly clean a home.
But the work ethic didn’t always work in Homejoy’s favor.
Late in 2014, when the company was already showing signs of trouble, Homejoy CEO Cheung posted a job titled “Dear Future Homejoy Engineer” on Hacker News, the Y Combinator news site. In the ad, she described how many employees were still in the office on Christmas Eve, playing 1990s pop music and fiddling with the algorithms. One Internet commenter called it “the most depressing job ad I’ve ever read.”
Luongo described the organization as “flat.” Executives and managers mixed with the rank and file, and rolled up their sleeves like everyone else. Adora Cheung opted to sit among employees, rather than in a swanky corner office. Many former employees described Adora as charismatic and accessible, and shared stories of her determination to right the ship and avert disaster, like picking up cleaning shifts when an alternative could not be found.
Requesting anonymity, one longtime employee said the founders grew overconfident and stopped paying much attention to their more experienced colleagues, including the executive team, after they raised the second round of funding. This person said that the founders had technical backgrounds (Aaron majored in chemical engineering in college and Adora previously worked as a product manager), but very little experience with customer support or home services.
“Advice often wasn’t heeded,” the former employee said. “There was a great deal of arrogance, especially after they secured the money.”
In one case, the source said that the founders were looking for ways to cut costs, and opted to scale back the customer service team. It was expensive to hire people to answer the phones and listen to customers’ issues from 8 a.m. to 8 p.m. across multiple time zones. Despite the advice of higher-ups, the founders decided to cut staff. That led to an uptick in wait times and rants from irate customers via social media.
To their credit, the founders did recruit more experienced executives in the final year, notably Kim Spalding, a former Starbucks’ vice president who now runs the home services team at Google. But several former employees said it was already too late as the requisite budget to right the ship was in short supply.
One employee described Spalding as “trying to [avoid] an iceberg when the ship’s already in motion. It’s hard to change direction.”
Worker Retention and The Problem of ‘Control’
Customer retention wasn’t the only stumbling block for Homejoy in the final years. It was a persistent struggle to retain high-quality cleaners.
For starters, Homejoy did not pay cleaners as much as some of its competitors, including Handy, a rival startup based in New York. Homejoy and Handy would both advertise for cleaners using similar methods, like a Craigslist ad or by recruiting from professional maid services, but Handy would routinely offer better rates in many cities (or at least the appearance of better rates if a candidate did not read the fine print).
After Homejoy took a commission, a cleaner would typically pocket $15 an hour. In many cases, that barely covered transportation costs and resulted in harrowing articles about cleaners who worked at all hours and could barely make ends meet.
It wasn’t until Homejoy’s final year in business that the company shifted focus to the experience of the cleaner. According to sources, client services started experimenting with new payment models, such as paying a higher hourly rate to cleaners who worked a certain number of hours or were able to retain customers.
If Homejoy did manage to recruit a new cleaner, it wasn’t always clear whether they were any good. Unlike Uber, which requires a drivers’ license, cleaners need more intensive training to do the job properly. It was a constant struggle for Homejoy to determine which cleaners were up to the task, and to keep the good ones loyal.
One limitation was the classification issue. In order to avoid litigation, Homejoy needed to meet IRS guidelines, which distinguish between employees and independent contractors. One of the biggest factors is the level of control exerted over a worker.
As a result, Homejoy hosted training sessions to guide cleaners through the basics in dozens of cities, but city managers could not mandate attendance. The company refrained from enforcing a dress code or uniform; they did not develop a list of tasks that a cleaner needed to complete.
Homejoy’s general approach to maintaining quality was to try out cleaners for a period and then boot them off the platform if they didn’t work out. Frequent issues included a lack of reliability, such as last minute cancellations, or too little availability. (Homejoy encouraged, but did not mandate, a minimum of 30 hours per week.) Given their independent contractor status, cleaners would not be formally terminated; they would simply stop receiving booking requests through the system.
Last-minute cancellations were a particularly vexing issue for the customer and client services teams, as Homejoy could typically only find a replacement about 15 to 20 percent of the time, according to a former employee. The problem was that in many of the largest cities, Homejoy didn’t reach a critical mass of cleaners. With short notice, it was difficult for most cleaners to travel across town to make an appointment on time. In these circumstances, it was not uncommon for local employees to volunteer to do a cleaning, including Adora on Thanksgiving Day.
“I used Homejoy from the early days,” said Colin Anawaty, 33, a startup worker who resides in Austin. “But I started looking for alternatives when I got a series of cancellations. I saw the writing on the wall.”
Cleaners who excelled would sometimes strike independent relationships with clients who wanted to see them again, but found it difficult to book a repeat visit via the Homejoy app. This often resulted in a pay increase, and some cleaners even attracted enough new clients to start their own small cleaning businesses.
Homejoy’s only recourse against this threat, known as disintermediation, was to stop working with cleaners who attempted to recruit customers. But it wasn’t always clear whether cleaners were seeking out independent relationships. Only late in the game did Homejoy attempt to make it easier for customers to request the same cleaner, which could have alleviated the problem or nipped it in the bud.
Lessons for the On-Demand Economy
Several of the former employees said they regretted that Homejoy didn’t engage more deeply in the national debate about the on-demand labor marketplaces. Some companies are pushing for a murky middle category that is not yet defined in the United States, but does have some precedence abroad. One idea is a dependent contractor, a model that gives workers more protections, including a severance package if terminated.
Instead, Homejoy’s top executives largely stayed silent on these issues.
Other experts say the fundamental issue with Homejoy wasn’t the founders’ public stance on labor economics and policy. A bigger question is whether the on-demand economy model, which has experienced some success with rides and odd jobs, even make sense for cleaning.
Cleaners require a higher level of training, and most can make more money by running their own business. Given the constraints of the 1099 model, it’s challenging to train workers to deliver a consistently high-quality experience.
On the customer side, it requires a level of trust to allow a stranger into your home, often unaccompanied, for several hours. For those who are willing to take the risk, every home is different and each customer will have a varied set of expectations.
“We were trapped between being accountable to our customers and not being able to take much responsibility for quality of service,” said Zietsman. “We couldn’t properly train our cleaners to meet fixed standards without fear of a legal backlash.”
Owyang, an analyst who has kept a close watch on this sector, said it remains to be seen whether tech-enabled cleaning services can succeed. But it’s possible that Handy and TaskRabbit can pick up where Homejoy left off, he said. Before it folded, Homejoy had introduced some additional services to tack on to a visit, like carpet cleaning, which could have netted them some much-needed cash flow.
In the near future, he said that technology behemoths like Google and Amazon might move into the space in a big way. Google scooped up much of Homejoy’s talent, including its former COO and a slew of engineers, which may signal a growing interest in cleaning and other home services.
Even so, Owyang predicted it would likely be years before a new, third worker classification is introduced to address the changes reverberating through the market, if ever.
Until then, he said, “many of these on-demand startups are in limbo.”
This is a one of a series of deep dives on subjects that illuminate issues concerning the future of the workplace. On November 12–13 in San Francisco, Medium and O’Reilly Media will co-host a conference on these issues called Next: Economy. You can find out more about it here.