In 2013, Gagan Biyani was working as what techies like to call a “growth hacker” at the ride-hailing company Lyft—his job was to help launch its service in Los Angeles, Lyft’s first market beyond San Francisco—when a childhood friend, Neeraj Berry, asked him if he’d be interested in applying Lyft’s blueprint to another type of service.

Cooking up loyalty: in January 2015, sprig CEO Gagan Biyani added a $10-per-month unlimited delivery option to his ondemand restaurant. “it’s better to increase a customer’s lifetime value,” he says, “than it is to increase your marketing spend” finding new customers.Photo: Noel Spirandelli

Biyani, a cheerful entrepreneur who liked starting companies, didn’t see why not. Three weeks later, he and Berry and two friends quit their tech jobs to found Sprig, an on-demand virtual restaurant that let consumers order dinner from a limited menu of healthy, homestyle dishes via its app. Biyani was CEO.

Biyani wasn’t the only entrepreneur at the time who saw opportunity in the Uber-ization of things. (Only Lyft alums like Biyani perceived the trend as a Lyft-ification.) Although Lyft and Uber had yet to grow into the disruptive and highly valued transportation companies they are today, Uber, in particular, had already shown promise. Leaked documents from that time revealed that the startup had managed to squeeze $213 million in revenues from $1 billion in gross volume, and investors who had missed out on ride sharing were now scrambling to apply the same system—connecting an army of independent contractors with customers through an app—to other services. Laundry. Liquor shopping. Weed delivery. Dog walking. Car washing. And, of course, eating. Every chore could now be a button on a smartphone: Press it, and problem solved. Venture capitalists who believed that Uber and Lyft’s model was widely applicable poured billions of dollars into this new convenience economy. Collectively, these startups became known as “Uber for X,” or the on-demand economy.

Sprig’s cofounders got to work on a Monday, and by Wednesday, they delivered 40 orders. They moved pieces from the board game Settlers of Catan around a map of San Francisco in order to keep track. Two months later, as business continued to boom, Biyani had developed more sophisticated tracking technology, but he also had an epiphany about “Lyft for food”: It would never work.

–– ADVERTISEMENT ––

While most licensed drivers can handle getting someone from point A to point B, freelance chefs working out of their own kitchens have numerous opportunities to fall short. One of Sprig’s cooks forgot to put the harissa chicken in half of the to-go boxes along with the roasted broccolini. Another, accustomed to cooking plate by plate in a restaurant, couldn’t replicate the taste of his dishes when he had to prepare them to serve 100 people. Sprig had recruited some talented cooks—in those early days, a sous-chef at George Lucas’s Skywalker Ranch and another who worked for San Francisco celebrity chef Michael Mina moonlighted for Sprig—but because each chef was operating his or her own kitchen, Sprig couldn’t order ingredients or operate as efficiently as more centralized operations. “We were losing a lot of money on every meal,” Biyani says, even though he was charging as much as $15 when some competitors were charging as little as $8.

Biyani realized that he couldn’t rely on independent contractors in the way that allowed Lyft and Uber to grow without ever hiring drivers or buying vehicles. Sprig’s first three staff hires were chefs.

In the past several months, the rest of the tech world has come to understand what Biyani discovered almost immediately: The on-demand economy is more complicated than merely applying a clever business model to different service sectors. None of the many startups that adopted Uber’s business model has managed to make it work as magnificently as Uber. Instacart (“Uber for groceries”) has had to reconfigure its staff and change its pricing to try to make its business work. Zirx (“Uber for valet parking”) shifted focus from consumers to businesses and will live another day. Other startups with less consumer appeal or operations prowess have simply shut down. “As eventually happens with any popular category, there is a phase shift from crowding to culling,” says Simon Rothman, a VC at Greylock Partners who has invested in both Lyft and Sprig.

Some pundits deem it the “on-demand apocalypse.” But what’s going on here is not so much the thinning of an oversaturated market as its maturation. On-demand companies use their networks and mobile technology to achieve a competitive advantage (and their traditional rivals are catching on quickly). But delivering food, it turns out, is not the same as dispatching cars. And providing child care is different from delivering food. This should not be a surprise: One can learn from a successful business model, but copying it verbatim almost never yields a similarly stellar result.

The businesses that once were happy to be lumped together in the hot trend are now realizing that, to survive and thrive, they can’t just be tech companies. “We’re not techies,” says Miguel Zabludovsky, cofounder of Slate, which was once thought of as “Uber for maid service.” “We’re cleaning 130, 140 houses every day.”

Beyond Point A To Point B

When Anand Iyer hails an Uber, he doesn’t really care who picks him up. All that matters is that he gets to his destination in one piece. Frankly, it’s a low bar. When Iyer hires someone to care for his 3-year-old daughter, Ava, however, his feelings are exactly the opposite. He wants someone who’s nurturing and responsible. Getting his kid back intact is expected. Child care is not a commodity service that can be performed by just anyone.

Training pays: on-demand babysitting app Trusted’s average gig lasts four and a half hours. “it’s unlike a quick grocery, valet, or car service,” says CEO Anand Iyer, who says he recoups the cost of his rigorous employee screening in the first two weeks of employment.Photo: Noel Spirandelli

So last year, when Iyer cofounded Trusted, an on-demand babysitting service in San Francisco, he didn’t even try to compete on price, the way that Uber and Lyft do with traditional taxi and car services. Instead, he hired employees who he knew he could count on. Before Trusted sitters begin work, they must undergo an extensive in-person interview, three reference calls, and a training session. “The value has to be 10 times better than anything else out there,” Iyer says, citing Trusted’s $25-per-hour fee. “It’s not like, ‘Let’s just try to do the best job we can in matching people.’ ”

Child care may be the most extreme example, but other services also require a surprising amount of training and standards. Last year, Instacart realized that it had to hire some of the people who would select, bag, and deliver groceries from the supermarkets with which it’s partnered.

Not just anyone, it turns out, can pick a ripe avocado from the produce department—and then pack an order for delivery so that the avocado arrives in a pristine state. There is no GPS for perfect produce. “This is something that people need to be trained on and coached on, on a regular basis,” says Instacart founder and CEO Apoorva Mehta. “And that’s how we succeed as a company, by making sure that it’s a business that customers want to use over and over again.”

A Fare Alternative

A few years ago, Marcela Sapone was juggling business school with errands as the on-demand economy boomed around her. Even as an overtaxed grad student in Boston, she didn’t think all of these supposedly convenient apps made her life that much easier. They were “kind of a novelty,” she says. “Many of us don’t need whatever we’re pressing the button for within the next 10 minutes.”

Hello, convenience: Hello Alfred coordinates its clients’ to-do lists and completes their errands each week. It cares less about on demand than getting things done behind the scenes. “We are the ultimate customer-service machine,” says CEO Marcela Sapone.Photo: Emiliano Granado

Instead, she wanted a service that would coordinate and complete the items on her to-do list, “so you don’t even have to pick up the app,” she says. Hello Alfred, which she cofounded in 2014, was originally thought of as a kind of meta-Uber, an on-demand service to manage all the other on-demand services one might use. The tech blog Valleywag lampooned Hello Alfred as the height of Uber for X absurdity, but it actually works much differently. Customers receive a personal assistant for $128 per month, and that person makes sure that the house cleaning, grocery shopping, and laundry gets done every week. With Sapone’s realization that “on demand” wasn’t as important as “easy,” Alfred removed itself from a reliance on Uber’s business model, which depends upon the unit economics of spontaneous, one-off transactions.

The company schedules employees so that they can serve all of its clients in an area on one route. Because Hello Alfred delivers convenience rather than instant gratification, it doesn’t need to recruit armies of independent contractors to stand by in case demand picks up (which Uber and Lyft accomplish through expensive tactics such as referral fees of up to $1,000). Hello Alfred, which boasts a 70% retention rate among its customers, has built service routes in 82 zip codes in New York, Boston, and San Francisco.

Other startups are also realizing that customers are a lot more flexible about when a task is completed. Just because the order comes via an app doesn’t mean that it needs to be done right now. The grass-mowing service Lawn Love and cleaning service Slate have adopted similar autopilot models.

Similarly, the dawning realization that Uber’s business plan only really works for ride sharing means that companies are getting more creative with how to make money. Instead of relying on a single transaction—like a fare—to make their businesses work, courier services Postmates and DoorDash can get restaurants and other merchants to pay for steering a customer their way. Postmates’ partners pay a fee of about 20% of an order in exchange for preferred placement.

Subscriptions are also emerging as a way to create both predictable revenue and customer loyalty. In January 2015, Sprig introduced Sprig Go, which grants customers who pay $10 per month the right to skip the typical $2 delivery fee on every order. One-third of Sprig’s customers now order three times as many meals as nonsubscribers, and they’re more likely to recommend the service to a friend. “If you wanted a great customer,” Biyani says, “you would describe them in those ways.”

A Virtually New Paradigm

Katherine Ryder, founder and CEO of a women’s health care startup called Maven, never intended to build an on-demand company. Instead of app-ifying an existing service, she was trying to resolve an issue that had never been adequately addressed in the first place. At the time, many of her friends were pregnant with their first children and struggling to make sense of prenatal care. “[Lack of] access and misinformation were the big problems,” she remembers.

Long-term healthy: “In the beginning, we won’t grow as fast as an ondemand laundry business, because we require more trust,” says Maven CEO Katherine Ryder of her telehealth startup. “But our market is enormous, and there are so many more potential sources of revenue.Photo: Emiliano Granado

Her idea was to develop a network of providers, like Uber, but without the last barrier that makes on-demand services such a challenging business: There would be no doctor running crosstown to take your case at the touch of an icon. An early Uber employee, Oscar Salazar, had already helped launch “Uber for doctors,” which would arrange an old-fashioned house call and cost as much as $200 per hour. Ryder used videoconferencing to connect patients with a network of mental health, reproductive health, and wellness professionals. Instead of booking an appointment at a medical office—which often involves planning days or weeks in advance, travel, and taking time off work—women who had questions about their pregnancy or general health could book a virtual appointment with a provider online. Providers pay a fee to Maven in exchange for using its software. When Maven finally launched, in April 2015, appointments started at $18 for 10 minutes. “The whole point is to make it as frictionless as possible,” Ryder says.

She reduced logistics and marketing expenses by selling Maven to businesses instead of just individuals, pitching companies on providing pregnant employees unlimited access to the service as a benefit. This summer, Maven will launch another subscription option for young women who are starting college. For $300 per year, their parents can purchase entrée to Maven’s network of health professionals so that whether they have the flu, need a birth-control prescription, or face depression, they’ll find help on Maven. “We’re not spending all of our time engaging customers in one-off transactions,” she says. “We’re delivering care in a way that it should be delivered, which is in a subscription, peace-of-mind, premium service.”

How An Uber Idea Grows Up

Today Sprig delivers thousands of meals every day in San Francisco and Chicago, within 20 minutes, on average, of when its customers placed the order. Biyani’s startup looks nothing like the Lyft-inspired company his childhood friend once proposed. Operations and the business model have been optimized for food; Sprig is a delivery-only restaurant. Maven, meanwhile, is a telehealth startup that provides employer benefits. Uber is a tool one uses to hail a ride to the airport. The three companies have so little in common that it is almost amusing that they’ve been grouped together as part of the same trend. After all, what defines an on-demand company in an era when technology and better customer service are a focus for every forward-thinking business? Do you loop in the $35 billion global services firm Price­waterhouseCoopers, which has an online portal that helps its consultants recruit independent workers for projects? What about $482 billion Walmart, for its in-store delivery services, or Chipotle and Starbucks, which have partnered with the courier Postmates to help them get burritos and coffee to their fans as quickly as possible?

Despite the doomsaying, “on demand” is not going away. What is really dying is “on demand” as a category. It’s not unlike what happened with the “Internet economy” of the late 1990s. All those celebrated “Internet startups”? We have a word for them today. We call them businesses.

An earlier version of this article incorrectly implied that Luxe had ended its consumer business.

A version of this article appeared in the July/Aug