Behind The Scenes At Homejoy, A Cleaning Startup That Says It's Really A Tech Company

If I were ranking startups based on how much I value their services, Homejoy would place pretty darn near the top — every month or so, one of their cleaners comes by my apartment and in two or three hours it becomes more sparkly than I've ever been able to make it. All for just 20 bucks an hour.

Homejoy's been growing quickly, too — it raised $1.7 million in seed funding from Andreessen Horowitz, First Round Capital, and others, and about 10 months after its official launch, the company says it now has a workforce of more than 50 employees in its San Francisco office. And yet … when I'm asked about exciting startups, Homejoy isn't the first one that comes to mind.

Some of my hesitation, I suspect, is related to the criticism leveled at a number of startups, that they're basically building services for lazy techies, rather than something practical for "the rest of us." (A criticism that's probably too big and complicated to be mentioned in passing in a post that isn't going to address the issue at all.) (Oops.) But there's also the question of whether Homejoy is even a tech company at all. Maybe it's just a cleaning service with a decent website? (The same question might be asked of competing startups like Exec, or of startups offering related services, such as laundry-focused Prim.)

Apparently this is something members of the Homejoy team were thinking about too, because they emailed me recently to suggest a guest column about why Homejoy really is a tech company. I made a counter proposal: How about I come to their office and see the technology in action? They agreed, and earlier this month, that's exactly what I did.

My tour guide was Mark Linsey, Homejoy's vice president of engineering. (He's second from the left in the photo above, which features Homejoy's five-person engineering team — and yes, they're hiring.) He told me that he was the company's first technical hire, joining as a consultant in January and then going full-time in March. Linsey, whose past work includes time as a technical program manager at Amazon and co-founding the social marketing startup Crowdbooster (which, like Homejoy, was backed by incubator Y Combinator), said he was convinced to stay by the size of the technical challenges that Homejoy was facing. Like Amazon, he said Homejoy's innovation is less about putting up a website for selling things (in this case home cleaning), and more about the backend technology.

"At Amazon, there's a whole software stack and product team for their warehouses and their internal logistics that they wouldn't be able to deliver the prices they do without," he said. Similarly, he claimed that with Homejoy, "It's like an iceberg. The customer-facing website is 5 to 10 perent of the whole."

So (to strain the metaphor) what technology is hidden under the water? For one thing, Linsey said Homejoy is "a very data-driven company." Matching the right cleaner to the right customer involves a lot of factors — not just how the cleaners and customers are rated, but also the routes that can maximize a cleaner's efficiency throughout the day, not to mention likely transit times in a given geography.

To illustrate this point, Linsey showed me the interface that Homejoy created for cleaners to identify exactly where they are and aren't willing to work. Originally, he said, cleaners identified their working areas based on zip code, but that turned out to be too broad (for example, many of the cleaners rely on mass transit, so in parts of the San Francisco Bay Area, many of them can only work near a BART stop). Now Homejoy gives them a tool where they can draw the exact, custom borders of their work area.

Linsey also showed me the "demand map" that Homejoy has created to display where its customers and cleaners are. The map works at a several scales, showing supply and demand across the entire United States (and in Canada, where Homejoy recently launched), or zoomed in to a specific geography — he showed me the Bay Area, which was crowded with multicolored pins. One color represented past customers, another showed upcoming appointments, and yet another stood for users who expressed interest in Homejoy but are outside the existing coverage areas.

The map is important for choosing new markets and finding new cleaners, Linsey said, recalling one occasion when the map revealed that Homejoy was starting to get a lot of jobs in the middle of the San Francisco Peninsula, an area where it didn't have many cleaners (they were more concentrated in San Francisco to the north and the Palo Alto/Mountain View area to the south), so that's where it focused its recruiting efforts.

The map, like a lot of Homejoy's technology, was built by Linsey and his team. There's also a custom CRM system for tracking cleaners, clients, and jobs, and a custom phone system, allowiing cleaners and clients to communicate without actually knowing each other's phone numbers. (The phone system is nicknamed Zoidberg, a nod to both the TV show Futurama and to Zoiper, the system that the company used before building its own)

None of this eliminates the need for a large customer service team. In fact, Homejoy let me listen in on a call with an initially unhappy client. On the call, a cleaner's car trouble meant that the company had to scramble to find a replacement (it helps that Homejoy pays some of its top cleaners to remain available on-call), and the aforementioned unhappy customer ended up getting a free cleaning — which seemed to to make up for a lot of the stress.

As for how Homejoy's approach is working for its cleaners, well, the ones I've spoken to have been pretty happy with the service. Special Projects Manager Marlo Struve told me that cleaners make between $12 and $15 an hour, and she noted that they have the freedom to determine where and when they're willing to work. She also sent me the following quote from cleaner John J. (the company doesn't identify cleaners by their full names): "Before [Homejoy] I had to solicit jobs myself and now Homejoy is like a household name. It's really picking up and I don't have to do as much [soliciting] as I used to [for my services]."

Linsey added that if I had stopped by the office in the company's early days, I would have seen a very different picture. The work that's now accomplished by his team's tools was originally done by co-founders/siblings Adora and Aaron Cheung, who worked out of their apartment and manually matched up cleaners and jobs. Homejoy's early approach, he argued, embodied Y Combinator founder Paul Graham's advice that entrepreneurs should "do things that don't scale" (advice that includes doing a lot of work manually at first, then building technology to automate the bottleneck).

Both Linsey and Struve said that without the technology and systems that were built after Homejoy's launch, the company could not have grown from one to 26 cities in just over six months. (And by the way, the list of markets where Homejoy is now available includes cities like Atlanta, Phoenix, and Tampa Bay, so maybe this isn't just for startup douchebags after all …) And for what it's worth, I was pretty impressed by what I saw — not just by the technology, but by Linsey's enthusiasm in outlining the problems Homejoy has already solved, and the ones that it still needs to face.

"There's a lot I think we can improve on in terms of gathering data," Linsey said. "I also think that Homejoy is not going to be exclusively a cleaning company forever — there are other services on the horizon."