‹ View all briefings
March 7, 2021

Beatdown on Aisle 5

The battle royale for the $12 trillion grocery market.

Photo by

Clarence Saunders was out on a limb. After all, there'd never been a store like it.

In the early hours of that September morning in 1916, perhaps the former grocery clerk took one last look at his creation. It certainly didn't bear a resemblance to any of the general stores at which he'd worked, starting at the age of 14. Gone were the countertops separating customers from the stockroom, so too, the barrels of wholesale goods — huge sacks of flour or drums filled with oats. 

Instead, wooden baskets greeted buyers upon entrance, and a parade of aisles demarcated the store. Rather than asking a clerk to pick up their order, Saunders wanted customers to shop for themselves. That meant passing the full range of his products — four times more than other stores — on the way to a cashier standing at the ready.

Would customers want to pick out products on their own? Saunders hoped so. But what excited him more was what his model unlocked from an economic perspective. Sure, there were some financial benefits to holding goods in bulk, but any advantages were overwhelmed by the sheer cost of labor. Every one of his old employers had relied on an army of clerks — an expensive necessity up to that point. By putting the onus of selection on the customer, Saunders could cut his staff and increase profits. That was without even factoring in the potential for his unique layout to encourage spending or lure sponsorship dollars from popular brands. 

With all that swirling through his head, numbers, narrative, and two decades of grocery experience, Saunders opened the door to Piggly Wiggly, the first self-serve grocery store in modern history. 

It was a smash hit. Within five years, Saunders rolled out 615 stores across the country, and by 1922, he took the company public. Not shy to toot his own horn, Saunders once proclaimed of his invention: 

One day Memphis shall be proud of Piggly Wiggly…And it shall be said by all men…That the Piggly Wigglies shall multiply and replenish the earth with more and cleaner things to eat.

While a Wall Street raid in 1923 would strip Saunders of ownership, the match had been lit. Over the next century, the grocery industry conformed to the Piggly Wiggly founder's vision of navigable aisles, individually packaged products, and self-service. 

As much as his store's actual design innovations, though, Saunders set in motion a mode of invention that continues to define the space. More than those that came before him, Saunders sought to fundamentally alter his chosen business's costs while introducing opportunities for financial upside. In Piggly Wiggly's case, such modifications focused on labor costs; today, many novel solutions add considerations like location, footprint, and throughput to the mix. 

In this briefing, we'll dive into the pitched battle taking place in the online grocery space, touching on dominant incumbents, novel upstarts, and intriguing infrastructural plays. We'll tackle: 

  1. A once-in-a-generation market opportunity
  2. Tricky economics
  3. Amazon and Walmart's advantages
  4. Farmstead and Picnic's clever models
  5. Vori and Fabric's infrastructure 
  6. Winning by saying no

Many of the current constraints would have been familiar to someone like Saunders. 

Market: The cart runneth over

If you alighted on a tech publication this week, there's a good chance you read something about neo-grocery. While a popular topic at the best of times, the last few days have been ludicrously active from a fundraising perspective:

  • Instacart raised an additional $265 million, valuing the business at $39 billion. That makes it the second most valuable private startup in the US, behind SpaceX. 
  • Rohlik, a Czech company, raised $230 million to fund European expansion. 
  • Flink, a German delivery business that leverages dark stores, raised $52 million.
  • Crisp, a Dutch startup that uses a fleet of EVs, raised €30 million with participation from Target's venture arm. 
  • Dija, an English company, confirmed it raised £20 million in December and is rumored to be closing in on an additional £100 million. 

To reiterate, over the past five days, as much as $750 million may have been allocated to grocery delivery startups. And while several of the companies mentioned above represent large, established entities, this is by no means a comprehensive list. Dozens of others are fighting for the same prize. 

What's driving this frenzy?

We can understand historic levels of investor interest as the result of three factors:

  1. Grocery's insane market size
  2. A once-in-a-generation shake out
  3. Winners' potential to capitalize on secondary opportunities 

Until scientists invent a way for humans to subsist on good vibes alone, the globe's 7 billion humans will need food regularly. Unsurprisingly, that means that the massive food and grocery market is large and likely to grow. The market was pegged at $11.7 trillion in 2019, growing at a 5% CAGR from 2020 to 2027. The largest regional markets are Asia and Europe, though the US boasts the most significant national spend, projected to be worth $1.7 trillion by next year. Of that figure, $800 billion is traditionally associated with grocery. China, India, Brazil, and Japan round out the top five by market size. 

Adapted from Statista

Over the coming years, growth is expected to be strongest in Asia, Africa, and the Middle East, with the mature markets of Europe and North America developing more slowly. 

That has not deterred investment in online grocery, particularly over the last year. Much of that is due to the coronavirus's destabilizing influence on the space. Stay-at-home orders have forced many to order groceries digitally for the first time, pulling forward penetration rates. 

In 2019, 5.1% of US grocery sales took place through an online portal. In 2020, that figure shot up to 6.6%. Similar patterns played out in other countries: sales jumped from 6% to 10.2% in France, 2.1% to 4.3% in Italy, 1.5% to 2.9% in Germany, and 8.1% to 12.4% in the UK. Asian countries also experienced an uplift, though because penetration was considerably higher pre-pandemic, this was less pronounced on a percentage increase basis. 

Adapted from Bain

Consultancy Bain described the shift as "[L]ike someone pressing the fast-forward button on the industry by several years."

Indeed, even in the six months between Bain's report and the start of 2021, figures seem to have changed considerably. A study from Retail Gazette suggested UK spend had risen to 14%; other countries are likely to have seen an uptick, too. Statista predicted that 51-57% of US consumers would have tried e-grocery by 2021. 

While some spending will evaporate in a post-vaccination world, as much as 45% of online orders could stick. BCG has branded the potential to capture this shifting spend as a "$2.5 trillion opportunity." Venture firms recognize there may never be a better time to win market share, explaining the sums raised. 

Along with the hefty prize already mentioned, some investors may point to online grocery businesses' ability to win ancillary markets as further justification. Because of the sector's tricky economics (details below), getting delivery right here suggests a capability to manage other supply chains. 

In discussing the potential of micro-fulfillment centers (MFCs) in grocery — essentially small warehouses optimized for delivery — one CEO in the space said

The question will come from the consumer: If I can get 20 to 30 items, why can't I get a shirt that I saw online at Nordstrom tonight?

Winning online grocery, or even a tiny part of it, would be reward enough. But should they wish to, those that dominate the category may find themselves best positioned to deliver all manner of goods: apparel, pharmaceuticals, at-home healthcare kits, and more. The opportunity is not just to siphon spend from Kroger but to compete with Amazon and become a piece of core, national infrastructure. 

First, though, companies will have to find a way to make the math work. As we'll see, that's far from a simple task. 

Economics: Shortchanged

Retail grocery stores have some of the most brutal economics of any business. Average operating margins are just 2-4%, even at scale. The largest grocery chains in the world remain governed by the gravitational pull of these percentages. Here are the operating margins of a few publicly-traded grocery companies: 

  • Kroger: 2.66%
  • Albertsons: 3.06%
  • Costco: 3.36%
  • Weis Markets: 3.97%

As you might expect, grocery thrives on volume as a result. Retailers are constructed to encourage large basket sizes and frequent purchases. This makes sense while the margin structure holds but becomes an accumulating liability under duress. While traditional costs include the products themselves, real estate space, labor, spoilage, and indirect expenses like advertising, offering online ordering opens up a slew of others. Among them:

  • Picking (selecting the items)
  • Packaging
  • Loading
  • Delivery

The real "cost" here is labor, precisely the expense Clarence Saunders sought to mitigate back in 1916. Ironically, in offering e-grocery services, many traditional retailers revert to a pre-Piggly Wiggly configuration, requiring an increase of "clerks" to select and fulfill customer orders. The benefits of self-service evaporate, and a century-old cost is added to the ledger. 

The result is brutal — retailers that provide online ordering without utilizing outside services or constructing bespoke facilities face operating margins of -15%. Even if analog stores require "click and collect," meaning customers select products in advance but retrieve it from a physical location, operating margins still slump to -5%. 

The harsh realities of these figures have provoked some to try alternatives, including partnering with third-party delivery services like Instacart or building specially made warehouses, centralized fulfillment centers (CFCs), or micro-fulfillment centers (MFCs). Alone, none are sufficient. 

Adapted from Statista

Services like Instacart bring operating margins up to a still grim -5%, unautomated warehouses have margins of -8%, automated CFCs log in at -7%, and automated MFCs reach -2%. Click and collect improves things somewhat with an automated MFC hitting 2% operating margins. While that's encouraging, transitioning to an MFC-only operating system would represent a massive expenditure likely to take years to implement, not to mention its unviability for environments with lower-order density. 

In sum, the outlook for online grocery looks bleak, with only the most expensive constructions making financial sense. That may make it the province of big players. 

Incumbents: Buying in bulk 

Though a relatively new space, online grocery has two clear incumbents: Amazon and Walmart. Both have taken the opportunity afforded by the coronavirus to press their advantages. 

Over the past year, Amazon added nine new warehouses in New York City, its largest market. A further two dozen or so were set up on the metropolis's outskirts. While the company will hope to use these facilities for various products, some portion will likely be allocated to grocery. By contrast, Walmart has no facilities within New York City proper, relying on warehouses in Pennsylvania.

Walmart, meanwhile, rolled out a Prime competitor, "Walmart+." The service offers members free delivery and additional perks for $98 per year. Amazon's service costs $119. That sell appears to be working — in its first five months of operation, Walmart+ attracted 7.4 million to 8.2 million members or roughly 13% of Walmart.com shoppers. As discussed in our recent Coupang piece, Amazon has reached 65% penetration with Prime. Given that Amazon launched Prime more than 15 years ago, Walmart's ability to get 20% of the way there is nothing to sniff at. 

Such innovations may partially explain how Walmart has overhauled Amazon as the number one e-grocer by number of transactions. According to TABS Analytics, a retail research product, Amazon held 27.1% of online food and beverage orders while Walmart garnered 30%. 

Given the study did not ask for transaction size, we should take relative positions with a grain of salt. The real takeaway is that Walmart and Amazon dominate the category already and are continuing to invest heavily. 

Neither will be willing to give up share without a real fight. Half of Walmart's annual revenue comes from grocery: of the $341 billion earned in the fiscal year of 2020, $191 billion was due to sales from the category. That means the company has an existential incentive to spend and win as much of the market as possible. Meanwhile, Amazon can subsidize low-margin endeavors through money-spinning divisions like AWS and clearly wants to own the space. 

How can new entrants compete?  

There may be some succor in returning to the TABS study. While traditional grocers saw their share of transactions drop from 15.7% to 15.3%, pure-play e-grocers saw a sharp rise. Grouped under "Other," companies like FreshDirect, NetGrocer, Peapod, and Instacart gained almost 9% share, rising from 7.9% to 16.7%. While the companies included are still relatively large (see: Instacart's valuation), it does indicate an ability for new players to change the landscape's dynamics and carve out a niche. Many such upstarts are taking creative approaches to alter the market's harsh economic realities. 

Insurgents: Discounts and markups 

Twenty-five years ago, Louis Borders, founder of Borders bookshop, unveiled his second act. Webvan sought to deliver groceries (and other products) within thirty minutes of the consumer's choosing. Produce came from the company's sophisticated Oakland Distribution Center, boasting five miles of conveyor belts, automated carousels, and real-time inventory information. Over the course of its short life, Webvan spent over $800 million before filing for bankruptcy. At the heart of its failure were upside-down economics. 

Today, more than a handful of insurgents in the online grocery space look a lot like Webvan, touting revolutionary tech and raising at rapid rates. The words "this time is different" strike fear into the hearts of rational investors, yet for many, that seems to be the primary argument. In some instances, they might have a point. 

New entrants use modern technology and innovative models to do one of three simple things: 

  1. Reduce the cost of real estate
  2. Reduce the cost of labor
  3. Increase product margins

Traditional grocery has long favored large, high-traffic locations. That real estate doesn't usually come cheap but could be justified because of the volume it brought in. The problem is that such sites are not optimized for an online ordering experience. As Saunders devised it, physical stores are built to increase time-in-shop. By making customers navigate serpentine aisles, grocery retailers encourage consumers to discover new products and add to their basket. That large, meandering format becomes a liability when grocery stores have to do the picking themselves, or even when outsourced to a third-party like Instacart. On average, Instacart couriers can complete two deliveries per hour. Paying for expensive real estate to justify that throughput makes little sense if viewed through the lens of online ordering. 

Insurgent online grocers forgo the in-person store in favor of CFCs and MFCs while increasing order volume. Rather than buying or leasing locations in popular neighborhoods, insurgents snap up facilities in less expensive areas optimized for selection and fulfillment speed. The result is a drastic reduction in real estate (and other) costs. 

Farmstead is a notable player in the space, primarily relying on CFCs. Described by its CEO as if "Whole Foods was rebuilt from scratch by tech founders," the company operates out of a 3,000 square-foot facility in the Mission District of San Francisco, along with additional "hubs" across the city. In theory, it is similar to the buildouts undertaken by UK grocer Ocado, German company Bringmeister, and Czech neo-grocer, Rohlik

The cost of launching these CFCs is $100K, according to Farmstead, while spinning up a new supermarket might cost $10 million. Thanks to its in-house technology, Farmstead's CFCs can each handle thousands of orders per day and reduce food waste by 3-4x. Profitability is apparently in sight, suggesting the model is viable. 

Miss Fresh is running a slightly different playbook on a much larger scale. With over 1,500 MFCs, the Chinese company has found a way to operate profitability at an efficiency level that makes larger companies blush. While the figures mentioned earlier suggest Instacart processes between 18 to 24 orders per courier per day, Miss Fresh handles 60 to 70. That represents a significant improvement over the US company as well as domestic super app Meituan, which manages 30 to 40. 

In an interview with GGVC, Miss Fresh COO Cecilia Sun details how this higher throughput influences thinking around economic yield per unit of real estate space (emphasis ours): 

[A] warehouse that is 200 square meters, we can deliver around three to four orders per square meter, that is efficient for per square meter. And per person, one delivery guy can deliver 60 to 70 orders per day, and that is efficient enough for delivery labor. So this efficiency is actually much higher than a store. So for stores, normally it is 100 RMB per square meter, that is the sales volume of store. So for us, one square meter is three to four orders. And our average order price is also around 100 RMB. So that is the 300 to 400 RMB per square meters, that is the average sales per square meter. And the delivery is 60 to 70 orders per day per person, and compared to Meituan (美团) which is around 30 to 40 orders.

Beyond those listed above, many of the latest online grocery stores are attempting a similar approach, including Gorillas (Germany), Flink (Germany), Weezy (UK), Dija (UK), Glovo (Spain), Cajoo (France), Getir (Turkey), Jüsto (Mexico), JD (China), and DoorDash (US). The related online convenience store space also relies on MFCs and brings in players like goPuff (US), Fancy (UK), and Zapp (UK). 

Beyond CFCs and MFCs bringing down the cost of real estate, they also reduce labor costs. This is another critical focus for neo-grocers, as it was for Clarence Saunders. By and large, the attempt to reduce personnel costs takes the form of two phases. 

In Phase 1, companies replace employees with contractors, stripping out associated benefits costs in making said transition. This model (with its associated societal baggage) has allowed gig economy businesses like Uber, Lyft, DoorDash, and Instacart to scale with comparative cost-effectiveness. 

In Phase 2, contractors are replaced by software and hardware robotics. 

In their embrace of automation, CFCs and MFCs have already entered Phase 2 in many respects. Rather than relying on an army of human pickers and couriers, facilities use robots to identify, select, package, and load various goods. Modern grocers are increasingly built for labor automation, while incumbent retailers catch up through partnerships.  Kroger is working with Ocado on two large CFC buildouts, H-E-B has tapped Swisslog for an MFC project, and Amazon is partnering with Dematic on a similar endeavor. Fundamentally, all are attempts to reduce the cost of labor. 

Reducing the cost of delivery will be the next challenge. Self-navigating robotics providers like Starship and Nuro seek to make human delivery drivers an anachronism, potentially removing another high cost. In time, we may see this become the norm. 

In the interim, though, some online grocers are finding other ways to reduce delivery expenses. Leading the pack is Picnic, an Amsterdam-based business. Rather than offering on-demand delivery, Picnic operates with a set route; essentially, acting like a bus rather than a taxi.

This decision reduces the number of trips Picnic makes, bringing down costs. The company passes savings onto the customer, allowing Picnic to sustainably price for the mass-market. (It has the attractive side effect of being considerably more environmentally-friendly, too, something the company augments by relying on EVs.) New routes are added only once a customer waiting list reaches sufficient length, de-risking expansion.

It seems to be a winning concept. At the start of 2020, Picnic reported 5% share of grocery spend in its most mature markets, a stake that is sure to have increased during the pandemic. In a nod to an earlier point made in this briefing, Picnic's co-founder Joris Beckers highlighted the company's ambitions beyond grocery, saying, "Food is our entry point, but we're building an e-commerce infrastructure." It will be interesting to see if others adopt Picnic's "Milkman" model. 

In addition to reducing real estate and labor costs, some insurgents seek to boost their bottom line by improving product margins. This is achieved by focusing on higher-margin items like convenience and specialty products or using emotional leverage to raise prices or levy additional fees.

As mentioned earlier, goPuff plays firmly in the convenience space. Rather than providing the full range of products available in a grocery store, the business has doubled-down on a comparatively limited range of items with higher margins. This is clear from its product expansion and M&A activity. Though it began by serving college kids, goPuff now offers "Home & Office," "Pets," and "Baby" supplies. None are aimed at that original market, but their appealing margins unite them. GoPuff would much rather sell you noise-canceling headphones, dog treats, or baby formula than a loaf of bread. Notably, the company acquired BevMo, an alcohol retail chain, for $350 million last November, bolstering its capabilities in another high-margin area. (Aside: the benefits of this limited SKU count trickle down, simplifying storage, fulfillment, and delivery.) 

Specialty grocers are making the same bet. Weee, an Asian and Hispanic vendor, is an example in this space, as is recent entrant Umamicart. Because these companies offer hard-to-find products, they're able to justify an elevated price and thus pad margins. The same calculation applies to high-end organic players like Good Eggs or Buffalo Market. Instead of pricing for the mass market, all four seek to make the math work by going upscale. 

Finally, insurgents levy new fees or increase prices through emotional means, effectively improving product margins. The clearest example of this is Zero, a neo-grocer that promises a plastic-free experience. The emotional and ethical appeal of that sell allows the company to markup individual products. Consumers can restore price-parity by signing up for membership. Jupiter makes a similar appeal, though less directly. Rather than an anonymous gig worker, customers are served by the same person each week, providing intimacy and connection.  The company will hope emotional appeal and white-glove service justify weekly membership fees and higher prices. 

Conspicuously missing from attempts to improve margins? Private label products. While Amazon has been aggressive in recognizing the profits it can realize by home-rolling basic brands, neo-grocers have, seemingly, yet to follow suit. In time, expect many to release their own lines — GoPuff Diapers are on the horizon. 

Infrastructure: In a food fight, sell napkins

The best-positioned companies in the grocery space may not sell food at all. As both modern and legacy businesses look to eke out efficiencies, next-level infrastructure companies may prosper. These providers offer a range of solutions to grocers, including advertising support, purchasing data, inventory analytics, waste reduction, and fulfillment solutions. In particular, the following five companies are at the heart of this part of the ecosystem: 

  1. Instacart, a provider of labor, advertising, and data. 
  2. Dumpling, a provider of labor with the potential to provide advertising and data. 
  3. Vori, a product to streamline wholesale ordering and reduce waste. 
  4. ShelfEngine, a product to improve inventory management and reduce waste. 
  5. Fabric, a provider of automated fulfillment services. 

At its core, Instacart is a "picks and shovel" play. The company provides a service to grocery retailers, fulfilling a core customer desire (delivery) at a lower cost than stores might manage themselves. Over time, the company will hope that technology allows them to reduce the costs of delivery further and partially explains why they are investigating partnering with retailers on MFC buildouts — if Instacart can replace humans for robots, they can pass savings to grocers while keeping a little extra for themselves. 

The actual picking and delivery service might be the least exciting aspect of the Instacart business, however. Beyond on-demand labor, Instacart provides two other core services to grocery retailers: advertising and shopping data. 

As Instacart has grown in popularity, its app has become the default place to begin a grocery purchase for more than 5.5 million users. That has presented a powerful advertising opportunity, similar to that earned by Amazon. When users search for a product ("cereal"), Instacart controls which brands surface first ("Special K"). That's pivotal, as consumers rarely scan more than a handful of options in making a buying decision. Proctor & Gamble executive Jon Moeller said the following: 

The assortment that a consumer exposes themselves to in an e-shopping trip is much lower [than expected]. Very few shoppers go to the third and fourth, much less the fifth and sixth pages of the search. So the question becomes what brands land on the first and second page of a search.   

Appearing on the first page is something for which the big brands are willing to pay. Instacart seems to recognize the opportunity, highlighting it as an area of future investment and aggressively hiring Amazon ad executives. Beyond this revenue stream, Instacart sells retail analytic data to brands, helping them understand who is buying what products. 

Dumpling offers a similar value proposition to consumers and grocers but with an entirely different structure. Instead of relying on an army of 1099s, Dumpling sells tools to couriers that want to operate as their own business. Grocery stores still effectively outsource their delivery, but rather than the customer ordering from Instacart, they do so from "Leila" or "Samuels Family Shopping" — small businesses leveraging Dumpling's software stack. 

This allows Dumpling to provide similar services for grocery stores while side-stepping employee classification issues and justifying higher prices (it feels much more rewarding to support a small business owner than to buy from a faceless corporation). In addition to a 5% platform fee, which Dumpling receives, individual shopping businesses add their own fees. Some of these reach as high as 25%, meaning consumers pay an extra +30% to have their order delivered. For reference, Instacart charges a 5% platform fee and a variable delivery fee, usually $3.99.

Vori sells operational efficiency to grocery retailers, making it easier to order from suppliers. With much of wholesaling done over phone, fax, and via paper catalogs, there's considerable room for improvement. While a streamlined digital experience should save time off the bat, platforms like Vori may be able to provide invaluable pre-purchase data. As stores look to improve margins, cutting down on waste and better understanding inventory levels could prove useful. 

Shelf Engine takes waste reduction a step further. By utilizing point-of-sale data alongside real-world information like "school schedules, local events, holidays, and weather," Shelf Engine helps retailers make more intelligent buying decisions. The company has so much confidence in its ability to forecast effectively that it promises to absorb any losses incurred from food waste. Many incumbent grocers are likely not applying this level of intelligence to their internal operations; partnering with someone like Shelf Engine may offset some of the costs of going online. While neo-grocers may be building some of this in-house, there's space for a dedicated offering that benefits from the scale of information processed. 

The most holistically valuable solution may be something like Fabric, a provider of automated fulfillment services. Rather than building a new core competency from scratch, grocers can partner with Fabric and spin up state-of-the-art CFCs or MFCs in relevant areas. Robots select and package items, software intelligently queues orders, and humans pick them up and deliver. Fabric can provide this service by retrofitting existing stores, kitting out a warehouse from scratch, or providing fulfillment as a service. The company's chief commercial officer described it as a kind of AWS for produce: 

It is a pure services model...You pay for capacity: how many totes you're taking up in my system, plus the volume of orders and the configuration of the order.

As alluded to earlier, others like Dematic, Ocado, and Swiss Log play in this space. 

It will be intriguing to watch which external capacities are eventually brought in-house and which carve out independent uses in the long-run. By and large, selling napkins in a food fight seems like an excellent game to play. 

Winners: The customer is probably wrong

"The customer is always right." 

If any maxim describes retail's overriding ethos, it is this pat phrase. 

Ironically, if insurgents steal share from Amazon and Walmart, it may be thanks to their ability to contract this aphorism. That will be particularly true for those that can creatively curb consumer discontent, framing limitations as benefits. Winners may be those companies that find a way to tell customers they are probably wrong. 

Companies will want to disabuse consumers of a few erroneous notions related to online ordering. Notably:  

  • "I should have an infinite selection of products available." 
  • "I shouldn't have to pay extra for anything." 
  • "I shouldn't have to wait more than thirty minutes."

As alluded to in the quote from the P&G chief, endless assortment is not as important to consumers as we might have thought. Rather than an infinite selection of products, we will see some winning neo-grocers focus on providing a targeted, differentiated selection with reasonable product margins. Weee does not try to offer every staple good or superstore SKU — it has a clear, narrow remit that earns the company a portion of grocery spend and gives it the latitude to price at a premium. Similarly, goPuff found an ideal beachhead to focus its higher-margin convenience store efforts: college students. While it didn't entirely do away with the need to visit the grocery store, goPuff recognized it could capture a reasonable percentage of spend with a limited SKU count. 

In both cases, these companies turned what might initially seem to be a disadvantage into a selling point, part of a unique value proposition. 

The second batch of winners will find a way of getting customers excited to pay more for delivery. There is no reason beyond unrealistic expectation that the added labor of selection and transportation should cost less, but that doesn't make adding a premium any less difficult. Though not a neo-grocer, Dumpling has found a brilliant way of messaging spending +30% by capitalizing on the wave of malcontent for abusive gig economy practices. The company outlines an ethical argument for the consumer and suggests greater personalization and community connection. Both Zero and Jupiter make similar claims, the former focusing on the moral pleasure of shopping sustainably, the latter highlighting customized service. All three companies are relatively immature — it's unclear which, if any, will thrive. But whichever player finds a way to sell higher prices as a perk should has a chance of making the math work. 

Finally, consumers will need to learn that sometimes, you have to wait. Many of the category's most recent entrants tout their speed, boasting about their ability to deliver groceries within 10 to 15 minutes. While this approach might work if allied with some of the strategies mentioned above, it will be tricky to square with a high-SKU, mass-market proposition. 

Picnic has found a solution by operating within temporal limitations. As noted, the company only opens up new geographies when demand has accumulated to a sufficient level. This is just one example. Picnic also requires customers to place their orders by 10 pm the day before delivery. This gives operations time to buy the produce, eliminating waste. Lastly, the company only offers a limited range of delivery times. 

While other companies might see these restrictions as drawbacks, they feel very much a part of Picnic's sell, fortifying messaging around sustainability, community, and affordability. In time, we may see others riff on this theme by requiring customers to pick up from designated points within a neighborhood (perhaps an MFC), removing a convenience by packaging it around similar themes.  

As Bezos once noted, consumers are "divinely discontented." We want more, cheaper, faster. In time, we should achieve all three of those things. As fulfillment is further automated, as delivery becomes the remit of robots, as software improves inventory planning and reduces waste, we may reach a juncture where virtually any product can be delivered to us within minutes, at a price lower than we could imagine. But until that is possible, grocery is a question of consumer tradeoffs. Savvy businesses will choose their inconvenience and build value around it. 


Saunders tried again. After Wall Street speculators took Piggly Wiggly from him, Saunders created "Keedoozle," an automated grocery store. In many respects, it resembled the retrofitted stores and MFCs of the present era. Upon arrival, customers received a key, which they inserted into slots that sat beneath their chosen items. Rather than picking it up themselves, the Keedoozle system logged their order by punching a hole into a strip of ticker tape. At checkout, customers presented the tape, which prompted workers behind the scenes to load produce onto a series of conveyor belts. Once they'd paid, customers could pick up their packaged items. 

It was a spectacular flop. Though impressive, Saunder's system was convoluted and prone to mistakes. Customers ended up with the wrong products and didn't enjoy the experience of shopping. Saunders tried yet again with a similar concept called "Foodelectric" but it met the same fate. 

The grocery industry has looked for ways to upend its margin structure for more than a century, beginning with Piggly Wiggly and continuing through to today's modern, automated MFCs. In Amazon and Walmart, the space has two deep-pocketed goliaths with the will and balance sheets to spend their way to success, taking the seeds of Saunders' ideas and making them palatable and profitable. 

Whether others can do the same remains an open question. With venture capital flooding the market and ingenious new models picking up momentum, there is some hope. Whatever the eventual outcome, the battle for grocery seems to have just begun. 

____

My thanks to Post Market, Mike Duboe, Dave Ambrose, Cristina Berta Jones, and Sid Jha for sharing their thoughts on the space. 

Clarence Saunders was out on a limb. After all, there'd never been a store like it.

In the early hours of that September morning in 1916, perhaps the former grocery clerk took one last look at his creation. It certainly didn't bear a resemblance to any of the general stores at which he'd worked, starting at the age of 14. Gone were the countertops separating customers from the stockroom, so too, the barrels of wholesale goods — huge sacks of flour or drums filled with oats. 

Instead, wooden baskets greeted buyers upon entrance, and a parade of aisles demarcated the store. Rather than asking a clerk to pick up their order, Saunders wanted customers to shop for themselves. That meant passing the full range of his products — four times more than other stores — on the way to a cashier standing at the ready.

Would customers want to pick out products on their own? Saunders hoped so. But what excited him more was what his model unlocked from an economic perspective. Sure, there were some financial benefits to holding goods in bulk, but any advantages were overwhelmed by the sheer cost of labor. Every one of his old employers had relied on an army of clerks — an expensive necessity up to that point. By putting the onus of selection on the customer, Saunders could cut his staff and increase profits. That was without even factoring in the potential for his unique layout to encourage spending or lure sponsorship dollars from popular brands. 

With all that swirling through his head, numbers, narrative, and two decades of grocery experience, Saunders opened the door to Piggly Wiggly, the first self-serve grocery store in modern history. 

It was a smash hit. Within five years, Saunders rolled out 615 stores across the country, and by 1922, he took the company public. Not shy to toot his own horn, Saunders once proclaimed of his invention: 

One day Memphis shall be proud of Piggly Wiggly…And it shall be said by all men…That the Piggly Wigglies shall multiply and replenish the earth with more and cleaner things to eat.

While a Wall Street raid in 1923 would strip Saunders of ownership, the match had been lit. Over the next century, the grocery industry conformed to the Piggly Wiggly founder's vision of navigable aisles, individually packaged products, and self-service. 

As much as his store's actual design innovations, though, Saunders set in motion a mode of invention that continues to define the space. More than those that came before him, Saunders sought to fundamentally alter his chosen business's costs while introducing opportunities for financial upside. In Piggly Wiggly's case, such modifications focused on labor costs; today, many novel solutions add considerations like location, footprint, and throughput to the mix. 

In this briefing, we'll dive into the pitched battle taking place in the online grocery space, touching on dominant incumbents, novel upstarts, and intriguing infrastructural plays. We'll tackle: 

  1. A once-in-a-generation market opportunity
  2. Tricky economics
  3. Amazon and Walmart's advantages
  4. Farmstead and Picnic's clever models
  5. Vori and Fabric's infrastructure 
  6. Winning by saying no

Many of the current constraints would have been familiar to someone like Saunders. 

Market: The cart runneth over

If you alighted on a tech publication this week, there's a good chance you read something about neo-grocery. While a popular topic at the best of times, the last few days have been ludicrously active from a fundraising perspective:

  • Instacart raised an additional $265 million, valuing the business at $39 billion. That makes it the second most valuable private startup in the US, behind SpaceX. 
  • Rohlik, a Czech company, raised $230 million to fund European expansion. 
  • Flink, a German delivery business that leverages dark stores, raised $52 million.
  • Crisp, a Dutch startup that uses a fleet of EVs, raised €30 million with participation from Target's venture arm. 
  • Dija, an English company, confirmed it raised £20 million in December and is rumored to be closing in on an additional £100 million. 

To reiterate, over the past five days, as much as $750 million may have been allocated to grocery delivery startups. And while several of the companies mentioned above represent large, established entities, this is by no means a comprehensive list. Dozens of others are fighting for the same prize. 

What's driving this frenzy?

We can understand historic levels of investor interest as the result of three factors:

  1. Grocery's insane market size
  2. A once-in-a-generation shake out
  3. Winners' potential to capitalize on secondary opportunities 

Until scientists invent a way for humans to subsist on good vibes alone, the globe's 7 billion humans will need food regularly. Unsurprisingly, that means that the massive food and grocery market is large and likely to grow. The market was pegged at $11.7 trillion in 2019, growing at a 5% CAGR from 2020 to 2027. The largest regional markets are Asia and Europe, though the US boasts the most significant national spend, projected to be worth $1.7 trillion by next year. Of that figure, $800 billion is traditionally associated with grocery. China, India, Brazil, and Japan round out the top five by market size. 

Adapted from Statista

Over the coming years, growth is expected to be strongest in Asia, Africa, and the Middle East, with the mature markets of Europe and North America developing more slowly. 

That has not deterred investment in online grocery, particularly over the last year. Much of that is due to the coronavirus's destabilizing influence on the space. Stay-at-home orders have forced many to order groceries digitally for the first time, pulling forward penetration rates. 

In 2019, 5.1% of US grocery sales took place through an online portal. In 2020, that figure shot up to 6.6%. Similar patterns played out in other countries: sales jumped from 6% to 10.2% in France, 2.1% to 4.3% in Italy, 1.5% to 2.9% in Germany, and 8.1% to 12.4% in the UK. Asian countries also experienced an uplift, though because penetration was considerably higher pre-pandemic, this was less pronounced on a percentage increase basis. 

Adapted from Bain

Consultancy Bain described the shift as "[L]ike someone pressing the fast-forward button on the industry by several years."

Indeed, even in the six months between Bain's report and the start of 2021, figures seem to have changed considerably. A study from Retail Gazette suggested UK spend had risen to 14%; other countries are likely to have seen an uptick, too. Statista predicted that 51-57% of US consumers would have tried e-grocery by 2021. 

While some spending will evaporate in a post-vaccination world, as much as 45% of online orders could stick. BCG has branded the potential to capture this shifting spend as a "$2.5 trillion opportunity." Venture firms recognize there may never be a better time to win market share, explaining the sums raised. 

Along with the hefty prize already mentioned, some investors may point to online grocery businesses' ability to win ancillary markets as further justification. Because of the sector's tricky economics (details below), getting delivery right here suggests a capability to manage other supply chains. 

In discussing the potential of micro-fulfillment centers (MFCs) in grocery — essentially small warehouses optimized for delivery — one CEO in the space said

The question will come from the consumer: If I can get 20 to 30 items, why can't I get a shirt that I saw online at Nordstrom tonight?

Winning online grocery, or even a tiny part of it, would be reward enough. But should they wish to, those that dominate the category may find themselves best positioned to deliver all manner of goods: apparel, pharmaceuticals, at-home healthcare kits, and more. The opportunity is not just to siphon spend from Kroger but to compete with Amazon and become a piece of core, national infrastructure. 

First, though, companies will have to find a way to make the math work. As we'll see, that's far from a simple task. 

Economics: Shortchanged

Retail grocery stores have some of the most brutal economics of any business. Average operating margins are just 2-4%, even at scale. The largest grocery chains in the world remain governed by the gravitational pull of these percentages. Here are the operating margins of a few publicly-traded grocery companies: 

  • Kroger: 2.66%
  • Albertsons: 3.06%
  • Costco: 3.36%
  • Weis Markets: 3.97%

As you might expect, grocery thrives on volume as a result. Retailers are constructed to encourage large basket sizes and frequent purchases. This makes sense while the margin structure holds but becomes an accumulating liability under duress. While traditional costs include the products themselves, real estate space, labor, spoilage, and indirect expenses like advertising, offering online ordering opens up a slew of others. Among them:

  • Picking (selecting the items)
  • Packaging
  • Loading
  • Delivery

The real "cost" here is labor, precisely the expense Clarence Saunders sought to mitigate back in 1916. Ironically, in offering e-grocery services, many traditional retailers revert to a pre-Piggly Wiggly configuration, requiring an increase of "clerks" to select and fulfill customer orders. The benefits of self-service evaporate, and a century-old cost is added to the ledger. 

The result is brutal — retailers that provide online ordering without utilizing outside services or constructing bespoke facilities face operating margins of -15%. Even if analog stores require "click and collect," meaning customers select products in advance but retrieve it from a physical location, operating margins still slump to -5%. 

The harsh realities of these figures have provoked some to try alternatives, including partnering with third-party delivery services like Instacart or building specially made warehouses, centralized fulfillment centers (CFCs), or micro-fulfillment centers (MFCs). Alone, none are sufficient. 

Adapted from Statista

Services like Instacart bring operating margins up to a still grim -5%, unautomated warehouses have margins of -8%, automated CFCs log in at -7%, and automated MFCs reach -2%. Click and collect improves things somewhat with an automated MFC hitting 2% operating margins. While that's encouraging, transitioning to an MFC-only operating system would represent a massive expenditure likely to take years to implement, not to mention its unviability for environments with lower-order density. 

In sum, the outlook for online grocery looks bleak, with only the most expensive constructions making financial sense. That may make it the province of big players. 

Incumbents: Buying in bulk 

Though a relatively new space, online grocery has two clear incumbents: Amazon and Walmart. Both have taken the opportunity afforded by the coronavirus to press their advantages. 

Over the past year, Amazon added nine new warehouses in New York City, its largest market. A further two dozen or so were set up on the metropolis's outskirts. While the company will hope to use these facilities for various products, some portion will likely be allocated to grocery. By contrast, Walmart has no facilities within New York City proper, relying on warehouses in Pennsylvania.

Walmart, meanwhile, rolled out a Prime competitor, "Walmart+." The service offers members free delivery and additional perks for $98 per year. Amazon's service costs $119. That sell appears to be working — in its first five months of operation, Walmart+ attracted 7.4 million to 8.2 million members or roughly 13% of Walmart.com shoppers. As discussed in our recent Coupang piece, Amazon has reached 65% penetration with Prime. Given that Amazon launched Prime more than 15 years ago, Walmart's ability to get 20% of the way there is nothing to sniff at. 

Such innovations may partially explain how Walmart has overhauled Amazon as the number one e-grocer by number of transactions. According to TABS Analytics, a retail research product, Amazon held 27.1% of online food and beverage orders while Walmart garnered 30%. 

Given the study did not ask for transaction size, we should take relative positions with a grain of salt. The real takeaway is that Walmart and Amazon dominate the category already and are continuing to invest heavily. 

Neither will be willing to give up share without a real fight. Half of Walmart's annual revenue comes from grocery: of the $341 billion earned in the fiscal year of 2020, $191 billion was due to sales from the category. That means the company has an existential incentive to spend and win as much of the market as possible. Meanwhile, Amazon can subsidize low-margin endeavors through money-spinning divisions like AWS and clearly wants to own the space. 

How can new entrants compete?  

There may be some succor in returning to the TABS study. While traditional grocers saw their share of transactions drop from 15.7% to 15.3%, pure-play e-grocers saw a sharp rise. Grouped under "Other," companies like FreshDirect, NetGrocer, Peapod, and Instacart gained almost 9% share, rising from 7.9% to 16.7%. While the companies included are still relatively large (see: Instacart's valuation), it does indicate an ability for new players to change the landscape's dynamics and carve out a niche. Many such upstarts are taking creative approaches to alter the market's harsh economic realities. 

Insurgents: Discounts and markups 

Twenty-five years ago, Louis Borders, founder of Borders bookshop, unveiled his second act. Webvan sought to deliver groceries (and other products) within thirty minutes of the consumer's choosing. Produce came from the company's sophisticated Oakland Distribution Center, boasting five miles of conveyor belts, automated carousels, and real-time inventory information. Over the course of its short life, Webvan spent over $800 million before filing for bankruptcy. At the heart of its failure were upside-down economics. 

Today, more than a handful of insurgents in the online grocery space look a lot like Webvan, touting revolutionary tech and raising at rapid rates. The words "this time is different" strike fear into the hearts of rational investors, yet for many, that seems to be the primary argument. In some instances, they might have a point. 

New entrants use modern technology and innovative models to do one of three simple things: 

  1. Reduce the cost of real estate
  2. Reduce the cost of labor
  3. Increase product margins

Traditional grocery has long favored large, high-traffic locations. That real estate doesn't usually come cheap but could be justified because of the volume it brought in. The problem is that such sites are not optimized for an online ordering experience. As Saunders devised it, physical stores are built to increase time-in-shop. By making customers navigate serpentine aisles, grocery retailers encourage consumers to discover new products and add to their basket. That large, meandering format becomes a liability when grocery stores have to do the picking themselves, or even when outsourced to a third-party like Instacart. On average, Instacart couriers can complete two deliveries per hour. Paying for expensive real estate to justify that throughput makes little sense if viewed through the lens of online ordering. 

Insurgent online grocers forgo the in-person store in favor of CFCs and MFCs while increasing order volume. Rather than buying or leasing locations in popular neighborhoods, insurgents snap up facilities in less expensive areas optimized for selection and fulfillment speed. The result is a drastic reduction in real estate (and other) costs. 

Farmstead is a notable player in the space, primarily relying on CFCs. Described by its CEO as if "Whole Foods was rebuilt from scratch by tech founders," the company operates out of a 3,000 square-foot facility in the Mission District of San Francisco, along with additional "hubs" across the city. In theory, it is similar to the buildouts undertaken by UK grocer Ocado, German company Bringmeister, and Czech neo-grocer, Rohlik

The cost of launching these CFCs is $100K, according to Farmstead, while spinning up a new supermarket might cost $10 million. Thanks to its in-house technology, Farmstead's CFCs can each handle thousands of orders per day and reduce food waste by 3-4x. Profitability is apparently in sight, suggesting the model is viable. 

Miss Fresh is running a slightly different playbook on a much larger scale. With over 1,500 MFCs, the Chinese company has found a way to operate profitability at an efficiency level that makes larger companies blush. While the figures mentioned earlier suggest Instacart processes between 18 to 24 orders per courier per day, Miss Fresh handles 60 to 70. That represents a significant improvement over the US company as well as domestic super app Meituan, which manages 30 to 40. 

In an interview with GGVC, Miss Fresh COO Cecilia Sun details how this higher throughput influences thinking around economic yield per unit of real estate space (emphasis ours): 

[A] warehouse that is 200 square meters, we can deliver around three to four orders per square meter, that is efficient for per square meter. And per person, one delivery guy can deliver 60 to 70 orders per day, and that is efficient enough for delivery labor. So this efficiency is actually much higher than a store. So for stores, normally it is 100 RMB per square meter, that is the sales volume of store. So for us, one square meter is three to four orders. And our average order price is also around 100 RMB. So that is the 300 to 400 RMB per square meters, that is the average sales per square meter. And the delivery is 60 to 70 orders per day per person, and compared to Meituan (美团) which is around 30 to 40 orders.

Beyond those listed above, many of the latest online grocery stores are attempting a similar approach, including Gorillas (Germany), Flink (Germany), Weezy (UK), Dija (UK), Glovo (Spain), Cajoo (France), Getir (Turkey), Jüsto (Mexico), JD (China), and DoorDash (US). The related online convenience store space also relies on MFCs and brings in players like goPuff (US), Fancy (UK), and Zapp (UK). 

Beyond CFCs and MFCs bringing down the cost of real estate, they also reduce labor costs. This is another critical focus for neo-grocers, as it was for Clarence Saunders. By and large, the attempt to reduce personnel costs takes the form of two phases. 

In Phase 1, companies replace employees with contractors, stripping out associated benefits costs in making said transition. This model (with its associated societal baggage) has allowed gig economy businesses like Uber, Lyft, DoorDash, and Instacart to scale with comparative cost-effectiveness. 

In Phase 2, contractors are replaced by software and hardware robotics. 

In their embrace of automation, CFCs and MFCs have already entered Phase 2 in many respects. Rather than relying on an army of human pickers and couriers, facilities use robots to identify, select, package, and load various goods. Modern grocers are increasingly built for labor automation, while incumbent retailers catch up through partnerships.  Kroger is working with Ocado on two large CFC buildouts, H-E-B has tapped Swisslog for an MFC project, and Amazon is partnering with Dematic on a similar endeavor. Fundamentally, all are attempts to reduce the cost of labor. 

Reducing the cost of delivery will be the next challenge. Self-navigating robotics providers like Starship and Nuro seek to make human delivery drivers an anachronism, potentially removing another high cost. In time, we may see this become the norm. 

In the interim, though, some online grocers are finding other ways to reduce delivery expenses. Leading the pack is Picnic, an Amsterdam-based business. Rather than offering on-demand delivery, Picnic operates with a set route; essentially, acting like a bus rather than a taxi.

This decision reduces the number of trips Picnic makes, bringing down costs. The company passes savings onto the customer, allowing Picnic to sustainably price for the mass-market. (It has the attractive side effect of being considerably more environmentally-friendly, too, something the company augments by relying on EVs.) New routes are added only once a customer waiting list reaches sufficient length, de-risking expansion.

It seems to be a winning concept. At the start of 2020, Picnic reported 5% share of grocery spend in its most mature markets, a stake that is sure to have increased during the pandemic. In a nod to an earlier point made in this briefing, Picnic's co-founder Joris Beckers highlighted the company's ambitions beyond grocery, saying, "Food is our entry point, but we're building an e-commerce infrastructure." It will be interesting to see if others adopt Picnic's "Milkman" model. 

In addition to reducing real estate and labor costs, some insurgents seek to boost their bottom line by improving product margins. This is achieved by focusing on higher-margin items like convenience and specialty products or using emotional leverage to raise prices or levy additional fees.

As mentioned earlier, goPuff plays firmly in the convenience space. Rather than providing the full range of products available in a grocery store, the business has doubled-down on a comparatively limited range of items with higher margins. This is clear from its product expansion and M&A activity. Though it began by serving college kids, goPuff now offers "Home & Office," "Pets," and "Baby" supplies. None are aimed at that original market, but their appealing margins unite them. GoPuff would much rather sell you noise-canceling headphones, dog treats, or baby formula than a loaf of bread. Notably, the company acquired BevMo, an alcohol retail chain, for $350 million last November, bolstering its capabilities in another high-margin area. (Aside: the benefits of this limited SKU count trickle down, simplifying storage, fulfillment, and delivery.) 

Specialty grocers are making the same bet. Weee, an Asian and Hispanic vendor, is an example in this space, as is recent entrant Umamicart. Because these companies offer hard-to-find products, they're able to justify an elevated price and thus pad margins. The same calculation applies to high-end organic players like Good Eggs or Buffalo Market. Instead of pricing for the mass market, all four seek to make the math work by going upscale. 

Finally, insurgents levy new fees or increase prices through emotional means, effectively improving product margins. The clearest example of this is Zero, a neo-grocer that promises a plastic-free experience. The emotional and ethical appeal of that sell allows the company to markup individual products. Consumers can restore price-parity by signing up for membership. Jupiter makes a similar appeal, though less directly. Rather than an anonymous gig worker, customers are served by the same person each week, providing intimacy and connection.  The company will hope emotional appeal and white-glove service justify weekly membership fees and higher prices. 

Conspicuously missing from attempts to improve margins? Private label products. While Amazon has been aggressive in recognizing the profits it can realize by home-rolling basic brands, neo-grocers have, seemingly, yet to follow suit. In time, expect many to release their own lines — GoPuff Diapers are on the horizon. 

Infrastructure: In a food fight, sell napkins

The best-positioned companies in the grocery space may not sell food at all. As both modern and legacy businesses look to eke out efficiencies, next-level infrastructure companies may prosper. These providers offer a range of solutions to grocers, including advertising support, purchasing data, inventory analytics, waste reduction, and fulfillment solutions. In particular, the following five companies are at the heart of this part of the ecosystem: 

  1. Instacart, a provider of labor, advertising, and data. 
  2. Dumpling, a provider of labor with the potential to provide advertising and data. 
  3. Vori, a product to streamline wholesale ordering and reduce waste. 
  4. ShelfEngine, a product to improve inventory management and reduce waste. 
  5. Fabric, a provider of automated fulfillment services. 

At its core, Instacart is a "picks and shovel" play. The company provides a service to grocery retailers, fulfilling a core customer desire (delivery) at a lower cost than stores might manage themselves. Over time, the company will hope that technology allows them to reduce the costs of delivery further and partially explains why they are investigating partnering with retailers on MFC buildouts — if Instacart can replace humans for robots, they can pass savings to grocers while keeping a little extra for themselves. 

The actual picking and delivery service might be the least exciting aspect of the Instacart business, however. Beyond on-demand labor, Instacart provides two other core services to grocery retailers: advertising and shopping data. 

As Instacart has grown in popularity, its app has become the default place to begin a grocery purchase for more than 5.5 million users. That has presented a powerful advertising opportunity, similar to that earned by Amazon. When users search for a product ("cereal"), Instacart controls which brands surface first ("Special K"). That's pivotal, as consumers rarely scan more than a handful of options in making a buying decision. Proctor & Gamble executive Jon Moeller said the following: 

The assortment that a consumer exposes themselves to in an e-shopping trip is much lower [than expected]. Very few shoppers go to the third and fourth, much less the fifth and sixth pages of the search. So the question becomes what brands land on the first and second page of a search.   

Appearing on the first page is something for which the big brands are willing to pay. Instacart seems to recognize the opportunity, highlighting it as an area of future investment and aggressively hiring Amazon ad executives. Beyond this revenue stream, Instacart sells retail analytic data to brands, helping them understand who is buying what products. 

Dumpling offers a similar value proposition to consumers and grocers but with an entirely different structure. Instead of relying on an army of 1099s, Dumpling sells tools to couriers that want to operate as their own business. Grocery stores still effectively outsource their delivery, but rather than the customer ordering from Instacart, they do so from "Leila" or "Samuels Family Shopping" — small businesses leveraging Dumpling's software stack. 

This allows Dumpling to provide similar services for grocery stores while side-stepping employee classification issues and justifying higher prices (it feels much more rewarding to support a small business owner than to buy from a faceless corporation). In addition to a 5% platform fee, which Dumpling receives, individual shopping businesses add their own fees. Some of these reach as high as 25%, meaning consumers pay an extra +30% to have their order delivered. For reference, Instacart charges a 5% platform fee and a variable delivery fee, usually $3.99.

Vori sells operational efficiency to grocery retailers, making it easier to order from suppliers. With much of wholesaling done over phone, fax, and via paper catalogs, there's considerable room for improvement. While a streamlined digital experience should save time off the bat, platforms like Vori may be able to provide invaluable pre-purchase data. As stores look to improve margins, cutting down on waste and better understanding inventory levels could prove useful. 

Shelf Engine takes waste reduction a step further. By utilizing point-of-sale data alongside real-world information like "school schedules, local events, holidays, and weather," Shelf Engine helps retailers make more intelligent buying decisions. The company has so much confidence in its ability to forecast effectively that it promises to absorb any losses incurred from food waste. Many incumbent grocers are likely not applying this level of intelligence to their internal operations; partnering with someone like Shelf Engine may offset some of the costs of going online. While neo-grocers may be building some of this in-house, there's space for a dedicated offering that benefits from the scale of information processed. 

The most holistically valuable solution may be something like Fabric, a provider of automated fulfillment services. Rather than building a new core competency from scratch, grocers can partner with Fabric and spin up state-of-the-art CFCs or MFCs in relevant areas. Robots select and package items, software intelligently queues orders, and humans pick them up and deliver. Fabric can provide this service by retrofitting existing stores, kitting out a warehouse from scratch, or providing fulfillment as a service. The company's chief commercial officer described it as a kind of AWS for produce: 

It is a pure services model...You pay for capacity: how many totes you're taking up in my system, plus the volume of orders and the configuration of the order.

As alluded to earlier, others like Dematic, Ocado, and Swiss Log play in this space. 

It will be intriguing to watch which external capacities are eventually brought in-house and which carve out independent uses in the long-run. By and large, selling napkins in a food fight seems like an excellent game to play. 

Winners: The customer is probably wrong

"The customer is always right." 

If any maxim describes retail's overriding ethos, it is this pat phrase. 

Ironically, if insurgents steal share from Amazon and Walmart, it may be thanks to their ability to contract this aphorism. That will be particularly true for those that can creatively curb consumer discontent, framing limitations as benefits. Winners may be those companies that find a way to tell customers they are probably wrong. 

Companies will want to disabuse consumers of a few erroneous notions related to online ordering. Notably:  

  • "I should have an infinite selection of products available." 
  • "I shouldn't have to pay extra for anything." 
  • "I shouldn't have to wait more than thirty minutes."

As alluded to in the quote from the P&G chief, endless assortment is not as important to consumers as we might have thought. Rather than an infinite selection of products, we will see some winning neo-grocers focus on providing a targeted, differentiated selection with reasonable product margins. Weee does not try to offer every staple good or superstore SKU — it has a clear, narrow remit that earns the company a portion of grocery spend and gives it the latitude to price at a premium. Similarly, goPuff found an ideal beachhead to focus its higher-margin convenience store efforts: college students. While it didn't entirely do away with the need to visit the grocery store, goPuff recognized it could capture a reasonable percentage of spend with a limited SKU count. 

In both cases, these companies turned what might initially seem to be a disadvantage into a selling point, part of a unique value proposition. 

The second batch of winners will find a way of getting customers excited to pay more for delivery. There is no reason beyond unrealistic expectation that the added labor of selection and transportation should cost less, but that doesn't make adding a premium any less difficult. Though not a neo-grocer, Dumpling has found a brilliant way of messaging spending +30% by capitalizing on the wave of malcontent for abusive gig economy practices. The company outlines an ethical argument for the consumer and suggests greater personalization and community connection. Both Zero and Jupiter make similar claims, the former focusing on the moral pleasure of shopping sustainably, the latter highlighting customized service. All three companies are relatively immature — it's unclear which, if any, will thrive. But whichever player finds a way to sell higher prices as a perk should has a chance of making the math work. 

Finally, consumers will need to learn that sometimes, you have to wait. Many of the category's most recent entrants tout their speed, boasting about their ability to deliver groceries within 10 to 15 minutes. While this approach might work if allied with some of the strategies mentioned above, it will be tricky to square with a high-SKU, mass-market proposition. 

Picnic has found a solution by operating within temporal limitations. As noted, the company only opens up new geographies when demand has accumulated to a sufficient level. This is just one example. Picnic also requires customers to place their orders by 10 pm the day before delivery. This gives operations time to buy the produce, eliminating waste. Lastly, the company only offers a limited range of delivery times. 

While other companies might see these restrictions as drawbacks, they feel very much a part of Picnic's sell, fortifying messaging around sustainability, community, and affordability. In time, we may see others riff on this theme by requiring customers to pick up from designated points within a neighborhood (perhaps an MFC), removing a convenience by packaging it around similar themes.  

As Bezos once noted, consumers are "divinely discontented." We want more, cheaper, faster. In time, we should achieve all three of those things. As fulfillment is further automated, as delivery becomes the remit of robots, as software improves inventory planning and reduces waste, we may reach a juncture where virtually any product can be delivered to us within minutes, at a price lower than we could imagine. But until that is possible, grocery is a question of consumer tradeoffs. Savvy businesses will choose their inconvenience and build value around it. 


Saunders tried again. After Wall Street speculators took Piggly Wiggly from him, Saunders created "Keedoozle," an automated grocery store. In many respects, it resembled the retrofitted stores and MFCs of the present era. Upon arrival, customers received a key, which they inserted into slots that sat beneath their chosen items. Rather than picking it up themselves, the Keedoozle system logged their order by punching a hole into a strip of ticker tape. At checkout, customers presented the tape, which prompted workers behind the scenes to load produce onto a series of conveyor belts. Once they'd paid, customers could pick up their packaged items. 

It was a spectacular flop. Though impressive, Saunder's system was convoluted and prone to mistakes. Customers ended up with the wrong products and didn't enjoy the experience of shopping. Saunders tried yet again with a similar concept called "Foodelectric" but it met the same fate. 

The grocery industry has looked for ways to upend its margin structure for more than a century, beginning with Piggly Wiggly and continuing through to today's modern, automated MFCs. In Amazon and Walmart, the space has two deep-pocketed goliaths with the will and balance sheets to spend their way to success, taking the seeds of Saunders' ideas and making them palatable and profitable. 

Whether others can do the same remains an open question. With venture capital flooding the market and ingenious new models picking up momentum, there is some hope. Whatever the eventual outcome, the battle for grocery seems to have just begun. 

____

My thanks to Post Market, Mike Duboe, Dave Ambrose, Cristina Berta Jones, and Sid Jha for sharing their thoughts on the space.