It’s become quite clear that running a so-called “on-demand” startup is no easy job. Already, several have succumbed to the difficult economics, increasing competition, and intricate operations, like home cleaning startup Homejoy, and laundry service Washio.
One such company that’s fighting to make it work is Instacart, a grocery delivery service from San Francisco. Founded in 2012, the startup lets customers order groceries from various stores it works with, and get them delivered to their door for a fee. And yet, despite being born in the age of Uber, it often draws comparisons to Webvan, an online grocery company that’s often deemed one of the biggest flops of the dot-com era.
But Instacart co-founder and CEO Apoorva Mehta rejects that comparison, of course. “Webvan failed 15 years ago, right? Lost of things have changed,” he said on Wednesday at the annual TechCrunch Disrupt tech conference in San Francisco.
According to Mehta, one massive technological change separates the two companies: smartphones. “[This] means that when you want to order your groceries, we can connect you to people who can shop for the groceries and bring them to your door, and this has never been the case before,” he said.
Get Data Sheet, Fortune’s technology newsletter.
Unlike Webvan’s bankruptcy—and even recent startups that have shut down—Mehta said his company will be “profitable” in the next 12 months. He clarified that he defines that as “cash flow positive,” which generally means that the company is taking in more money than it’s spending, though it doesn’t necessarily mean it’s profitable. (Fortune‘s Dan Primack discussed the startup’s economics with Mehta in March.)
So how does Instacart plan to achieve that? Like most companies promising to deliver some sort of service or product with the tap of a mobile app, the company’s revenue comes from the fees it charges customers and the fees it charges merchants. But it’s also banking on a third source: advertising.
In the case of companies like Instacart, those ads don’t look like traditional ads. Instead, a consumer packaged goods brand pays to have its products prominently displayed within Instacart’s app. In the brick-and-mortar retail world, it’s the equivalent of, say, end caps—those special product displays at the end of an aisle that heavily display that product or brand. In retail stores, consumer packaged goods companies pay to be featured there, or for any other more prominent type of display, for that matter. Instacart already has partnerships with Coca-Cola (ko), Unilever (ul), and Proctor & Gamble, among others.
Speaking of the company’s relationship with grocery stores, Mehta also added that Instacart works with them on more than one level. While customers mostly just know that the startup puts their products on its app and delivers orders from their stores, Instacart is also working to provide them with software tools to better track orders and manage their inventory. The challenge there, is that many of them are still using very old technology.
One grocery chain working very closely with Instacart is Whole Foods Markets (wfm), the Austin-based chain focused on healthy products. Instacart works with over 100 grocery partners, but Whole Foods is one of the few that has stores in all the markets the delivery service operates, said Mehta.
Asked if Instacart would sell to Whole Foods, which has invested in the startup, Mehta dismissed the idea, adding that “it just doesn’t make sense for us to even think about selling to a grocery store.” Moreover, contrary to popular assumption, Whole Foods isn’t the biggest source of order volume for the company, or at least not in all markets.
Another partnership that’s going well is with Publix, a grocery chain operating in the Northeast of the U.S. Publix sales through Instacart, which began earlier this summer, are growing by 12% week over week, according to Mehta.
But of course, not everything has gone smoothly for Instacart, especially in terms of personnel. For example, in March, the startup slashed the wages of its shoppers and delivery workers in some major markets like San Francisco and Los Angeles, further fueling industry concerns that delivery startups might work better in theory than in practice.
Mehta admitted that “when you grow so fast, obviously you’re gonna have adjustments that you’re gonna have to make,” adding that the wages it paid when it only served three market didn’t make economic sense by the time it had expanded to 15 markets, and beyond. Mehta also dismissed the idea that while employees working for Instacart’s corporate operations get great salaries and a flurry of Silicon Valley perks, contractors who shop and delivery the grocers aren’t treated as well. “We want to make sure that our shoppers are compensated fairly,” he said.
In December, the startup laid off a dozen recruiters, which again, Mehta explained away by saying that it was just part of the adjustments it had to make after a growth spurt. Within the next couple of months, Instacart plans to release its first report on the demographics of its employees, following a recent trend among tech companies by exposing how diverse (or not) its company is.
“I think diversity is something that we take very seriously at Instacart,” he said, adding that everyone as his company undergoes unconscious bias training. With that said, unconscious bias training’s effectiveness is increasingly being questioned.
But if there’s one thing Mehta said that’s difficult to argue against is that “the problem we’re trying to solve is very hard.”