Hi 👋 - While Amazon is raising grocery delivery prices, investors are throwing billions at startups offering ultrafast, cheap grocery delivery. Today, a look at GoPuff, one of the pioneers in the space. Thanks for reading.
If you’re finding this content valuable, consider sharing it with friends or coworkers. ❤️
For more like this once a week, consider subscribing. ❤️
Inflation Is Real
On October 25th, Amazon raised prices for Whole Foods grocery delivery, to the consternation of yuppies everywhere1. Two-hour delivery, formerly free for Prime Members, now costs $9.95. It also raised the price of one-hour grocery delivery to $14.95 from $9.95. Inflation is real.
According to Amazon, the delivery fee covers operating costs, including equipment and technology associated with grocery delivery, and helps keep grocery prices competitive. Prime members can still shop online and pick-up orders over $35 at Whole Foods for free. Amazon might be tipping its hand here.
There are two steps in fulfilling an online grocery order:
Picking & Packing: This is a fancy way of saying outsourced grocery shopping and refers to all in-store activity. When an order is placed online, a pick list (your grocery list) is generated at the nearest store. This list is given to a shopper who prepares the order by picking items off the shelf. Once the order is picked, the items are bagged and prepared for delivery (packing).
Delivery: Getting the packed groceries from a grocery store to your front door.
Amazon is a savvy capital allocator, world-class operator, and focused on the long-term. Because it’s analytically rigorous and shrewd, its new fee structure suggests that it struggled to make the unit economics of free grocery delivery work. Yet at the same time Amazon is raising grocery delivery prices, billions of venture capital dollars are flowing into a cadre of startups like GoPuff, Gorillas, JOKR, and Weezy promising cheap, ultrafast delivery (often 15 minutes) of convenience items and groceries. The space has attracted over $7 billion in investment the past three years2. At $15 billion, GoPuff’s most recent valuation is higher than the $13 billion that Amazon acquired Whole Foods for in 2017. The discrepancy between Amazon’s price increase and the ultrafast delivery fundraising makes you go, hmmm.
Grindin’
While Amazon doesn’t disclose details around the economics of grocery delivery, the financial performance of DoorDash and Uber Eats suggests that delivery is a challenging business. In the third quarter of 2021, Uber Eats approached EBITDA breakeven at $12.2 billion in gross bookings, its most profitable quarter to date. Nelson Chai, Uber’s CFO, said that while the core restaurant delivery business is profitable, investments in non-food delivery like booze, grocery, and pharmacy (or ultrafast delivery startup terrain) are driving the losses.
DoorDash’s profitability is slightly better, though by no means robust. Over the past year, it has operated with one percent EBITDA margins.
To be fair, DoorDash and Uber Eats are both in investment mode, depressing current profitability, and call out that profitability is higher in their more mature markets. Still, delivery is a grind. It’s operationally complex and the influx of funding for ultrafast delivery startups means competitive intensity is increasing. For example, in London, there are at least eight ultrafast delivery startups in operation3. There will be blood.
What is GoPuff?
GoPuff is one of the largest and oldest ultrafast delivery providers. Founded by Drexel University classmates Rafael Ilishayev and Yakir Gola in 2013, GoPuff began as an on-demand delivery service for hookah and other smoking products4. The company quickly pivoted to convenience items, like pints of ice cream, bags of chips, and rolls of toilet paper, but kept its logo and name.
GoPuff’s value proposition is convenience. It's attacking the gap between a larger supermarket and the bodega or corner store. To achieve marketplace liquidity, Ilishayev and Gola constrained their market by focusing on college students (similar to Facebook) and convenience. After establishing the business in Philadelphia, GoPuff expanded into cities with large student bodies like Austin, Boston, and Washington DC. Today, it has 10,000 employees, operates in over 1,000 cities, and has expanded beyond college students, who represent less than fifteen percent of customers5. It has also increased its selection from late night munchies to a wider array of food as well as baby care, booze, pet supplies, and over-the-counter medicines.
Vertical integration is central to GoPuff’s business model, allowing it to control the customer experience and optimize for speed. The company owns and operates its micro-fullment centers (MFCs), negotiates directly with brands, and custom-built its own warehouse management software. MFC layouts are optimized to minimize pick and pack time, which can happen in under two minutes6. Because all orders start at the same origin point, a GoPuff MFC, batching orders is easier. This is a critical advantage and a differentiator with delivery marketplaces DoorDash and Uber Eats, where orders originate from multiple pickup points. GoPuff’s traveling salesman problem is easier to solve, meaning its couriers can potentially have higher utilization and provide faster delivery.
GoPuff operates over 500 MFCs and is opening about 30 per month7. Customer proximity being critical to speedy delivery, these are typically located in city centers and densely populated neighborhoods, and serve a delivery radius of one to two miles. MFCs are about 8,000 square feet and hold roughly 3,500 SKUs. Systemwide, GoPuff works with about 20,000 SKUs8. Each unit is merchandised with a selection of local products, in addition to national brands like Ben & Jerry’s. A modest SKU count means GoPuff can concentrate its purchasing power and get discounts. Limited selection is a big difference compared to Amazon, which focuses on endless selection and offers 170,000 SKUs between Whole Foods and Amazon Fresh.
Within a given metro, GoPuff uses a fortification strategy to build its business. This means packing multiple MFCs into a city. Theoretically, higher MFC density increases brand awareness and decreases delivery times, helping the city get to profitability faster. This is a capital intensive strategy requiring leases, headcount, and inventory.
Operationally, the number of orders delivered per hour (ODH) that a courier can make is a critical metric. Today, this is between four and four-and-a-half GoPuff’s best performing markets, though based on an average delivery time of over twenty minutes, the company wide average is likely between two-and-a-half and three. Density of demand and density of MFCs critical to ODH. GoPuff’s goal is to reduce aggregate delivery times to 15 minutes from the low-twenties today, which would increase ODH (assuming there’s enough demand).
From a unit economics perspective, GoPuff makes its margin on products, not delivery fees. It charges a delivery fee of $1.95 per order and has gross margins in the high-forties. Comparing an equivalent basket of goods between GoPuff and Whole Foods, the groceries were ten percent cheaper at Whole Foods, but GoPuff had a lower total cost after accounting for delivery fees:
Economically, a GoPuff order breaks down into: grocery gross margin plus delivery fee less cost of delivery less cost of fulfillment less overhead and fixed costs (MFCs, demand generation). Assuming a forty-five percent gross margin, GoPuff generates about $18 on a $35 order. This $18 has to cover the cost of delivery, operations, customer acquisition, retention, and corporate overhead. In addition to delivery, the company also has nascent subscription and ads businesses, providing high-margin revenue streams.
The unit economics are there to build a sustainable business on. GoPuff will live or die based on the efficiency of its customer acquisition, its ability to retain customers, and the level of corporate overhead.
Your Margin Is My Opportunity
GoPuff says it is profitable and generating cash flow in all markets where it has been operating for at least eighteen months. Impressive, but without seeing GoPuff’s income statement, there are reasons to take this with a grain of salt. For example, executives from Sweetgreen and WeWork claimed to be profitable before IPO filing documents proved otherwise. Additionally, it's unclear how profitability is defined; EBITDA definitions are notoriously fickle. To be fair, GoPuff’s management talks about the importance of unit economics and profitability, which isn’t always the case with growth-obsessed startups.
It’s also unclear if GoPuff has a durable advantage in ultrafast delivery. The company’s vertical integration, know-how (bespoke warehouse management software, for example), and operational experience certainly give it a leg up. But the fortress strategy relies on MFC density and MFC density relies on capital, which is flooding into the space. Competition from startups and incumbents is mounting. In Europe, Uber Eats partnered with Carrefour to provide grocery delivery in Paris and DoorDash is mulling opening its own MFCs. Additionally, a good chunk of Manhattan lives within a mile from a Whole Foods:
When asked about the risk of ultrafast delivery during its Q3 2021 earnings call, Amazon CFO Brian Olsavsky responded:
It's a rapidly evolving space. And obviously, we're customer-obsessed, but we also are competitor aware. We like the business that we have. We have over 170,000 products that customers, Prime customers can get within 2 hours from Amazon Fresh, Whole Foods Market and other stores that participate with us in over 5,000 cities and towns. So we're well on our way to providing ultrafast delivery for things that require ultrafast or things like groceries and others. And we see that expanding. But there'll be room for multiple winners in the space. And as we say, you have to have a cost structure and a logistics network that will pay for the delivery over time. So we see it as part of an offering that we offer to customers that ranges from 2 days to 1 day to 2 hours or 1 hour in some cases. So we like to meet customers where they are and when they need things, and we're working on speed consistently.
Delivery is a hard business. Consistent, ultrafast delivery is even harder. Amazon’s recent price hike suggests that its cost structure and logistics network wasn’t adapted to provide free, fast grocery delivery profitably. Unlike Amazon, GoPuff’s business model is optimized for limited selection, convenience-based pricing, and speedy delivery. Results to date have been impressive, but it’s still day one for GoPuff. It will be years before we know if it has the cost structure and logistics network to pay for delivery over time.
If you’re finding this content valuable, consider sharing it with friends or coworkers. ❤️
For more like this once a week, consider subscribing. ❤️
More Good Reads
If you’re interested in learning more about GoPuff’s business model, unit economics, and expansion strategy, Harry Stebbings’ interview with co-founder Rafael Ilishayev is a good listen. Delivery is eating the world. Below the Line on Sweetgreen’s love-hate relationship with DoorDash and Uber Eats.
Myself included.
The Wall Street Journal, A New Market Emerged for Online Delivery: 10 Minute Groceries, November 1, 2021.
The Wall Street Journal, A New Market Emerged for Online Delivery: 10 Minute Groceries, November 1, 2021.
Overlooked by Alexanre Dewed, A Deep-dive into goPuff's Strategy to Conquer the Grocery Market, November 16, 2020.
GoPuff Bloc, A Peek Behind The Curtain: Gopuff’s Unique Business Model, September 3, 2021.
VC 20: The Twenty Minute VC, 20VC: GoPuff’s Rafael Ilishayev on How GoPuff Has Been EBITDA Profitable From Day 1; The Unit Economics Behind GoPuff, With Intense Competition What Happens To The Food Delivery Space & What It Takes To Launch, Grow, and Maintain New Markets, June 14, 2021. Unless otherwise notes, all financial and unit economics stats in this section are from this interview.
The CPG Guys, Delivering Immediate Needs with GoPuff’s Alina Bilger & Andy Berman, May 5, 2021.
Great one, Kevin. Very interesting space that requires really rigorous process for operational output. We have some in Toronto, like Ninja. Fewer SKUs and well placed MFCs means a model that can work - in very dense urban centres. Grocery delivery is not the same as meal delivery, the latter can get slightly better margins. At what point does the consumer choose to just go to a store i.e. is the marginal convenience of delivery worth the fee / somewhat inflated prices? Quality is also key.