The Gig is Up for Gig Economy Grocery

Welcome back to The Innovation Armory. I wanted to start off this piece by thanking all of my awesome subscribers who have helped me scale the newsletter to where it is today. Over the coming weeks, I plan to also launch a corresponding podcast series leveraging content from the newsletter. All free, like the newsletter. Stay tuned.

I am a newbie to Twitter (just made an account a couple days ago) and if you enjoy the newsletter content, would appreciate if you’d follow my new account @SamNatbony.

If you’ve enjoyed the newsletter, I would also greatly appreciate if you would share this article on LinkedIn, Twitter, Facebook, etc. For subscribers who share my content, I plan to provide exclusive additional content and perks like the opportunity to ask questions of founders for upcoming pieces. You can share at the below link:

This week’s post is on the rise of a new model in digital grocery delivery that leverages vertical integration and logistics optimization packaged in tech-enabled services model rather than the capital light approach of players like Instacart. I believe this new model has the potential to totally transform the grocery delivery segment and disintermediate gig economy players. Thanks so much to Ricardo Weder (CEO of Jüsto) and Alberto Menolascina (CEO of DIJA) for sharing your perspectives for this piece. This is a long post so if your email gets clipped at the bottom make sure to click unclip / visit The Innovation Armory to check out the full read.

My next piece will be on creator and micro-influencer networks powering the growing passion economy. Thanks in advance to Gustav Lundberg Toresson (CEO of Memmo) for sharing your perspective.

I hope you enjoy this piece and feel free to subscribe for future updates below:

Traditional grocery is a particularly challenging retail segment in which to operate. Many of the largest traditional grocery chains in the US command low margins due to high labor costs, inventory complexities, food spoilage costs and a greater propensity for consumers to order groceries online. Traditional grocers often earn operational margins in the 4-6% range excluding taxes, depreciation and interest expense:

This lower margin profile makes it difficult for grocers to accept incremental variable costs when expanding their distribution strategy into new channels including digital. I believe this dynamic poses a long-term threat to the core revenue streams in the business model of grocery delivery businesses like Instacart. Instacart primarily makes money in three different ways: commissions, service fees and product advertising.

Instacart is reliant on its grocer network to support its business model, but over the long-term, it is actually eroding margin for grocers, displacing their brands and bleeding value from potential new revenue streams. It not only reduces the significance of their grocery brand but actually can create negative affect brand as consumers associate negative ordering experiences like missing goods, late deliveries and melted frozen items with the grocer rather than with Instacart. Grocers own the local distribution (in the form of retail stores) as well as a network of connections with producers, CPG companies and suppliers. By choosing to sell on Instacart, top grocery chains are rent these assets to capture short-term increases in demand in ways that are will be eventually destructive to their own business. At scale, the commissions that Instacart charges are not scalable for grocers given their existing razer thin EBITDA margin profile. If Instacart were to lower commissions at scale, those costs to Instacart would most likely need to be absorbed by the end consumer in the form of higher delivery fees. Instacart is popular in more modern, urban city center areas, and while higher earning areas may be okay absorbing some higher service costs, these higher costs would not scale well outside of tier 1 cities. Even if consumers would be willing to accept these higher fees, I think any rational big box grocery chain would rather build out its own delivery fleet to capture a potentially higher margin revenue stream to augment its razor thin margins. Even the advertising revenue stream of a gig marketplace, like Instacart, is zero sum with grocer incentives because i) it is revenue traditional grocers could otherwise be capturing on their own native e-commerce sites and ii) it erodes the value of shelf space in their stores for which grocers have historically charged brands. Vertically integrated grocers that effectively transition into the digital era have the power to recapture this value from Instacart and build a more sustainable business model with margins way in excess of current grocery levels.

Instacart’s relationship with grocers is very much akin to the decisions by i) third party retailers as to whether to sell their goods through Amazon and ii) telecom / production companies’ decision to license their content to Netflix. The smartest and most forward thinking brands either stopped doing so early or never participated because doing so would change the basis of competition and erode their own brands and their connection with the end consumer over time.

Over the next ten years, traditional grocery brands and the next generation of digitally native grocers will similarly recapture value from Instacart by prioritizing best-in-class digital first capabilities and optimized logistics networks to re-assert their direct connection with their customers. Traditional grocers without core competencies in technology may be slower to make this transition. However, there are already a number of new digitally native and vertically integrated grocery players emerging who are well positioned. Their vertical integration will enable them to simultaneously improve traditional grocery margins, while also delivering value back to consumers in the form of lower prices by owning the connection between food producers and suppliers. 

To learn more about the digitally native grocery space, I sought out leaders building vertically-integrated grocery ecosystems. I first spoke with Ricard Weder, CEO & Founder of Jüsto, a Mexican online grocer that recently raised the largest Latin American Series A round of the past decade. Ricardo also served as former CEO & President of Cabify, the Madrid-based ride-hailing company that raised nearly $500M of capital. Next, I spoke with Alberto Menolascina, CEO & Co-Founder of DIJA, a UK-based digital first grocer that uses hyper-local fulfillment to delivery groceries in 10 minutes or less. Alberto was also a part of senior management at Deliveroo and REEF previously. Unlike traditional grocery, both businesses are digitally native and differentiate through best-in-class logistics networks and optimized consumer experiences. I found both conversations particularly insightful given their experiences working in the gig economy (in taxis and food delivery respectively) and their lessons learned from those experiences on how next generation grocers can recapture value from gig economy players like Instacart. Please enjoy my conversations with Ricardo and Alberto.

A Snapshot of My Conversation with Ricardo Weder (CEO of Jüsto)

SN: How did your experience in ride-hailing inform how you’d build and scale Justo?

RW: What we are building is completely different than ride-hailing but that doesn’t mean we can’t draw lessons from it. We are very different than the aggregators like Rappi and Cornershop. We have our own micro-fulfillment centers and we deliver same-day or next day. We own the whole value chain and benefit from all the data that is collected because we own the whole process from end-to-end. Something I learned from ride-hailing is that there are minimal barriers to entry assuming other players have access to cash and sponsors willing to burn that cash. That’s why I wanted to fully verticalize Jüsto so that we could build barriers to entry and drive towards profitability. From Cabify, I also learned the importance of leveraging data to drive differentiation and how to scale teams on a by-city basis in a fast way to capture share. 

SN: Could you in further depth about the benefits of vertical integration vs. the more capital light approach of sharing / gig economy players?

RW: If you analyze the margins in gig economy grocery, at the end of the day, they have all of the costs of a supermarket, plus their own costs. Sooner or later these costs will be paid by the user or the grocer and the user acquisition subsidies from the delivery or ride-hailing companies will end. This fee can represent up to a 15% markup and in Latin America, the number of consumers willing to pay this market vs. other markets is very low. Further, when you source from the supermarkets, you depend on their portfolio of products for innovation, which tends to be fairly static. Owning the sourcing element lets you interact with the consumer in a more meaningful way. Because you own the whole chain and the corresponding data, you can actually drive costs down for consumers instead of adding that 15%+ premium. Over the long-term, I believe most supermarkets will realize that the cost of not owning your own user is very high. The supermarkets that develop their own e-commerce channels will stop participating in ride-hailing companies. We are already beginning to see that. In Mexico, the largest traditional supermarket already recently stepped out of the aggregator ecosystem. To bring real innovation to the industry, you need to own the full system to drive system-wide efficiencies.

SN: Could you elaborate on some of the nuances of the food ecosystem in Latin America down from producers to consumer relative to other parts of the world. Where are the inefficiencies in the current system?

RW: In Mexico and many countries in Latin America, we have enormous central markets. Wholesalers come to central markets to sell their products. However, these markets are not really adding to the supply chain, especially for small and medium farmers. In the case of Jüsto, 20-30% of our providers are farmers who normally don’t sell in a supermarket because of the commercial conditions. Sometimes they are paid very late, like 180 days after contributing goods to the market, which is unsustainable for a smaller business owner. Plus, these central markets are highly informal and fragmented with a lot of intermediaries extracting value. I believe there is a huge opportunity still to optimize the distribution of food within Latin America. You see studies that show in the region that the waste of fresh food in the channel is up to 65%. Agents throw it away if it doesn’t have the right color and often the distribution is delayed and the food spoils. We saw these problems and wanted to improve them with technology and fair practices, which is why we are called Jüsto, which means fair in Spanish. 

SN: What are the different pockets your the logistics network that you have most innovated on to more effectively scale your model relative to competitors?

RW: We have a model of small and medium-sized micro-fulfillment centers. Depending on the transportation capabilities of the food producer, they either bring the food products to a micro-fulfillment center or to a more central drop-off and sortation distribution center. From there we distribute to the other micro-fulfillment centers. One of the largest costs that erodes profitability in the grocery sector is delivery and distribution costs. We have been able to really reduce these costs to 4-5% (half of traditional e-commerce) because we are very close to the users. As we grow and have more density of orders, we gain operating leverage and the data also tells us where to put our next micro-fulfillment center. In Latin America, the biggest barrier to ordering groceries online is that consumers don’t trust the quality of fresh food. That is our main differentiator by both having micro-fulfillment centers closer to the customer (keeping the food fresher) and by empowering smaller farmers to contribute high quality fresh products. Also, in Mexico, because we have central markets in big cities, the sourcing infrastructure is used to going through the central markets. These central markets are hubs because of the lack of cold chain capabilities currently outside of urban areas in Mexico. We intentionally built our logistics infrastructure by the central markets and close to highways so that we could tap into the existing selling patterns of local and smaller produce sellers. 

A Snapshot of My Conversation with Alberto Menolascina (CEO of DIJA)

SN: Could you discuss the opportunity for improvement you saw in gig economy grocery from a consumer perspective?

AM: These businesses tend to see lower retention from end customers related to missing items, substitutions, frozen products that melt and what I call, The Paradox of the Delivery Slot. If today is Thursday, you are going to order online so you don’t have to go to the store, but now the earliest date for a delivery slot is Monday. Now you have to bridge Thursday through Monday with supplemental grocery purchases until your order arrives. Many consumers are also upset about the long delivery windows where they often wait 2-3 hours at home they would have otherwise used in another way. Then you start unpacking another products in the fridge and you start by eating the products that are due to expire soon rather than eating the fresher goods and the food you want that night. After four days you don’t eat fresh, you eat what’s left like a student might or paradoxically, you go back to the store. You were meant to order online not to go the store and you end up having to do two replenishments.

SN: What technology and system design innovations have you leveraged to drive differentiation at DIJA?

DIJA’s logistics network is effectively a two-tiered distribution system with a central warehouse outside the city center and a distributed network of micro-fulfillment hubs very close to customers in densely populated areas with a radius of about 1 mile. This allows us to deliver in 10 minutes and we deliver primarily through electric scooters and small vehicles, which allows us to deliver faster than competitors. We have a selection that varies depending on the neighborhood between 2,000 - 4,000 SKUs. This SKU range is both the greatest weakness and strength of the model. We don’t have a huge variety of selection like a large store, but at the same time we offer a full grocery experience in terms of SKU coverage and can decide which food vendors and produce are in and which are out. We don’t outsource delivery to third parties because then we can’t hit our promised 10 minute delivery window. For us logistics is not just about efficiency, it is about labor law and making sure that you treat drivers in an ethical way. 

SN: Beyond back-end and system design innovations, how are you specifically improving the customer ordering experience?

Relative to other players, we are also very focused on the digital experience. If the promise to the customer is a delivery in 10 minutes, then it is important to correspondingly keep the order time down to 60 seconds or less. Some people believe you can’t order for a full week’s worth of food in 60 seconds. I want to flip that hypothesis on its head and re-orient customer expectations around ordering for just the day what they need or for meaningfully smaller periods of time than a week. People are accustomed to ordering for a week because the experience of ordering can be so cumbersome that you want to spread the return on your time investment over a week so you don’t have to go back into the application and order again. However, if you could place an order in a few seconds and receive in a few minutes, I believe customers would not want to order for a week at a time. We are striving to have an Instagram-like experience where every time that you open the app, it knows what you like and what you don’t. If you go the supermarket and you don’t have a pet, you wouldn’t repeatedly walk the aisle with pet food. Similarly, grocery apps that don’t leverage personalization should not waste order time by including or showing you products they know based on your prior clicks you should not like. Equally important is the digital ordering experience and making it as visual as when you go to the store.

Very often people ask me how you are going to be different from others. I don’t think that’s the right question. Grocery is grocery and at the end of the day it comes down to execution. Being better means having a system that is so scaleable that every cent that we save goes back to the customer in the form of lower prices. I see a lot of people now going crazy with models with massive subsidies and discounts for customers to join without refining the model. I don’t think this is a great way to approach a market. When Deliveroo won London against companies like Take it Easy, then Uber Eats came into the city and in 6 months gained almost 30% market share because the market was already educated by Deliveroo. At the end of the day, competitors that give away un-economical subsidies are education the market at their own cost and one day we will win over their customers. At the end of the day, the only thing that matters is pricing and the ETA that you promise. We are optimizing our execution strategy to hit those targets and don’t care if we lose a customer today who we will get back in 6 months with a better, more economical model. 

SN: Could you comment on the practice of competitors using aggressive user acquisition subsidies in this space and if you have contemplated a similar strategy? What other levers do you use to drive fast customer acquisition?

AM: Being better means having a system that is so scaleable that every cent that we save goes back to the customer in the form of lower prices. I see a lot of companies now going crazy with massive subsidies and discounts for customers to join without first refining the model. I don’t think this is a great way to approach a market. When Deliveroo won London against companies like Take it Easy, then Uber Eats came into the city and in 6 months gained almost 30% market share because the market was already educated by Deliveroo. At the end of the day, our competitors that give away un-economical subsidies are educating the market at their own cost and one day we will win over their customers. At the end of the day, the only thing that matters is pricing and the ETA that you promise. We are optimizing our execution strategy to hit those targets and don’t care if we lose a customer today who we will get back in 6 months when we will have an even better, more economical model. When you acquire customers that converted only because of discounts, the month after you have to acquire them again. The retention profile is much worse than for a customer who actually believes in the value proposition at the cost at which it can be delivered. Execution on the experience promised so that customers actually receive their order in ten minutes creates evangelists who will propagate their experience to their own network. If you think about Gorillas, they grew incredibly fast in Germany because the second best player in the space could only deliver the next day. Beyond subsidies, it can be very disruptive to promise virally fast and exceptional service levels and actually consistently deliver on that vision. In my mind that that is much more valuable than giving out a one-time voucher and acquiring lower quality customers with the promise of free food. 

SN: Given your plans to scale rapidly to hundreds more cities over the next couple of years, how are you prioritizing which regions to go after first?

AM:  There are two different strategies to city expansion: the Uber / Deliveroo and Door Dash / Just Eats models. Uber Eats in the US took over the market in densely populated urban city centers. For DoorDash, the urban city center has never been the main area of focus. Glovo was successful in Spain because they realized that in Spain, price is much more important than service levels. Even to provide savings of 50 cents to one euro, it was worth delivering 12 minutes later. That buys you time to stack orders and save on the delivery costs, much more than the level of discount that you are giving. With last mile logistics, you can play a lot with price and service level to create an optimal combination that meets the standard demographic of that market. In London, it is very important to deliver fast and can charge a markup, but in tier two cities that markup strategy may not be as effective. In our case, we are going to be focused on big cities at first, but then will go after surrounding areas with different and personalized logistics algorithms. 

SN: Given some capital intensity to setting up distribution centers, how do you iterate on and experiment with the model in cost effective ways?

AM: When you want to build a big business in this space, you also have to take on the burden of capital intensity. It is a matter of understanding what is the right form factor. We are testing different store sizes in each city we launch. Among our first ten locations, we tried ten different things. We will also treat new cities as opportunities to experiment with our hub models with regards to factors like parking spaces, street access, traffic density, warehouse layout and more so that our system design gets smarter as we scale further. When you look at the capex side, it is not too high if you don’t involve robotics and we layer on robotics we will save other costs. Ghost kitchens by contrast are very high initial capex because the kitchen equipment is more costly to buy and maintain. For a dark store, the biggest cost is a freezer and is less complex equipment-wise.

SN: Could you talk about the importance of culture and how you use culture to drive a competitive organizational advantage?

AM: The best thing you can do culturally as a leader is to go to your sites and speak with those packing goods and fulfilling and delivering orders to understand how their work should inform your standard operating procedures. They are experts at their jobs and they know what works and what doesn’t, what is intuitive and what will waste time. Career progression and development is also critical. If you hire people just to use them as one small piece of the small value chain without room for upward mobility, they will never provide valuable input. If they believe that the person who delivers our orders on a scooter can one day become the CEO of the company and they have the tangible opportunity to step up, they will provide great feedback and be incentivized to be greater employees and execute. It is important to create visibility to give feedback to employees in real-time. Hands on learning also is critical to our culture. In running such a complex system, our engineers regularly go to our warehouse to understand how they work and will sometimes code from our warehouses. If you don’t have the mentality to make your hands dirty and run the most basic procedures, you won’t be able to develop a technology that works. I tried to build a business from the office before and realized I was so distant from the execution that the difference between the vision and execution is not what I expected working behind my screen. It is about bringing a good group of people together and getting everyone (including executives, engineers, strategy folks) to understand that the real world is not their sofa. The warehouse, the people and the execution determine the profitability and success of the business. 

Vertically Integrated > Capital Light Models

A primary reason gig economy models gained favor amongst investors is due to their capital light approach, i.e. Uber doesn’t own any taxis, Instacart doesn’t own any stores or physical logistics infrastructure. In the case of grocery, I believe a vertically integrated, digital first approach with smart capital investment is the long-term winner in the space. Amazon drove outsized demand for its e-commerce platform largely due to massive backend logistics investment that enabled it to deliver the best customer experience and gave it real, physical barriers to entry. Similarly, in grocery, the most value will not accrue to software only players like Instacart. Rather, it will accrue to logistics and services companies that are able to drive outsized services demand for D2C delivery via technological innovation. For grocery, that means new players need to invest capital in micro-fulfillment centers to build their competitive moat and drive outsized delivery demand by being able to deliver faster, cheaper and better.

Vertical integration in grocery also has the potential to unlock some powerful data advantages as modern grocers sit at a valuable position between food suppliers / producers and consumers. In a world with Instacart, consumer data sits in a silo that is separate from producer data and the two never interface. There could be substantial efficiency improvements to allowing these data sets to talk to another:

Vertically integrated grocers who have their own digital ordering experience can leverage the power of consumer data combined with supplier data to create additional value for both consumers and producers. On the producer side, consumer purchasing and demand data can better inform local farmers on the quantity of crops they need to plant, how quickly they need to harvest and which organic products are trending in their area. On the consumer, side, this engagement with food growers unlocks the opportunity to offer bulk consumer discounts through group buying initiatives similar to those employed at Pinduoduo, which I discussed in The Manifesto of Social Commerce. Outside of financial and planning benefits, better coordination between consumers and growers can significantly reduce food waste and thereby help to decrease the carbon footprint of farming and help create a more circular food economy.

Separately, many gig economy business models have effectively created two separate classes of workers within companies whereby management and delivery workers / warehouse workers are paid highly disproportionately, with limited promise of upward mobility for blue collar workers on these platforms. I agree that senior management at Uber adds more value to the organization than any one driver and should be compensated appropriately. However, I believe the magnitude of that pay difference should at the very least be compressed enough to give effectively full-time drivers proper benefits. Because gig economy models earn their margin on top of all of the underlying costs of the asset they connect to the market (e.g. Instacart’s margin can only be paid if the traditional grocer earns gross margin on the product), many have tried to achieve higher margins by squeezing value from their contractor labor force. Even outside of the gig economy, Amazon has been criticized heavily for its warehouse worker practices. Companies DIJA and Jüsto that emphasize fair labor practices will have an organizational innovation advantage by treating all stakeholders in their organization fairly, including delivery workers, through upward mobility opportunities and a greater appreciation by management of the day in the life of even the lowest paid employees in the organization. When you are trying to hit narrow delivery windows, the operational excellence fostered by this alignment of incentives makes a difference at hyper-scale.

Psychology Meets Grocery: Choice Paralysis and Order Cadence

In Got FOBO? Try a Discovery Platform, I discussed how too much choice can itself be a friction in the purchasing process. Traditional grocery stores stock lots of SKUs of similar products with the goal of providing customers more choices, but often end up wasting time by making their stores more difficult to navigate, increasing their own inventory costs and creating negative customer sentiment through choice paralysis:

Using data to curate a smaller set of fresh and in-demand products may be preferable for some consumers in terms of reducing clutter that contributes to choice paralysis. Even for the long tail of grocery clients that have a very particular set of niche product preferences about a particular brand of product, many would trade off that choice for a handful of products if that meant unlocking the convenience of their full order arriving in less than 30 minutes. Even if some customers don’t want to make that tradeoff, the volume increase from a better delivery experience would likely outpace any volume loss attributable lower SKU count. This concern is further mitigated by the fact that most branded products that would fit into this long tail of customer preferences have a longer shelf life and could be bought separately in bulk from other channels while still utilizing a digital first grocer for convenience for fresher products.

More importantly, if many customers would trade convenience for SKU diversity, grocers could meaningfully improve margins and cash flow generation:

The first two means of margin improvement are relatively self-explanatory, but the third requires a bit more explanation. Currently, traditional grocers view additional SKUs as an opportunity to differentiate from other grocers through a unique, bespoke product set. This mentality gives food producers (especially of branded products) relative leverage over grocers in regards to pricing terms. However, a digital first, dark store grocer has limited SKUs and can actively shape which consumer products are winners and losers in their local markets through which products they choose to prioritize in their limited product allotment. This flips the traditional grocer dynamically on its head where grocers take it as a given which consumer products are popular and therefore which they must stock. Consumers will gravitate to grocers that can offer the best digital customer experience and fulfill / deliver the fastest and it will become table stakes for the best consumer brands to penetrate these channels. I believe this power shift will give these grocers a pricing advantage as they command better terms from food suppliers. This provides these grocers a) an opportunity to reduce their variable COGS and b) an ability to create a flywheel of more consumer volume demand by passing these price savings back onto customers. This is similar to the pre-sales distribution power dynamic shift I discussed in Future Proofing Digital Education.

In enterprise software, often an intuitive user interface that feels natural to consumers is touted as a significant competitive advantage. Great UIs reduce user friction and create a positive emotional response by channeling innate human tendencies to create product workflows that are better molded to our inherent human psychology. Apple, for example, initially differentiated its hardware products from first generation computers with UI software that was intuitive, like drag and drop and animated icons like the trash can. Similarly, in retail, next generation grocers can differentiate by structuring their networks and applications to be more consistent with a more intuitive ordering cadence of smaller daily orders, which would help create a more delightful rather than stressful ordering experience:

The first generation of digital grocery in the form of Instacart tried to apply the traditional grocery model of buying food for a full week at a time to a digital first format. This made sense when grocers had to invest substantial time and resources into an offline grocery trip. However, this is not an intuitive ordering cadence in the context of the internet. Unlike traditional retail goods, a customer’s demand for any particular food they may want can shift significantly on a daily basis. You may crave pizza today but want to make a salad tomorrow. It is not ideal to have to lock oneself into a food preparation schedule based on your cravings today that may be different in the future. That’s why many people who order groceries online find themselves running out before each meal to spot purchase other ingredients they need to create a particular dish they want on any given night. We think in the present when it comes to our eating, not one week ahead, so it is much more intuitive and better channels our innate psychological tendencies to create a food ordering system that allows consumers to get the groceries they want today. First generation online grocery didn’t develop this way because the traditional grocery ordering model was improperly translated online and was combined with a capital light approach that led to the insufficient buildout of local micro-fulfillment to sustain these systems from a logistics perspective. Further, I also believe there was too much of an emphasis on growing average order value vs. growing average order frequency per user. The former is a better growth driver in businesses where users have a natural inclination to order less frequently. The latter is better in industries like food where users would love to order as close to real-time as possible if afforded the opportunity.

Customer Acquisition Costs Trump Customer Education Costs

Alberto discussed how user acquisition subsidies may be ineffective if businesses are acquiring lower quality users that will churn easily. This got me thinking about an interesting framework that could be a useful way to think about subsidy specific marketing spend in the context of a technology businesses. As it relates to subsidies and discounts, there is a meaningful distinction between customer acquisition costs and customer education costs.If business spends prematurely and aggressively in a growing sector on subsidy-based sales and marketing initiatives before it has refined its business model, it may not be subsidizing user acquisition growth, but rather be educating customers about a market trend that one of its competitors will be better positioned to capitalize on. Further, when taken to an extreme, aggressive spend too early can lead to negative word of mouth marketing if quality issues consistently repeat.

It is certainly more of an art than a science figuring out which bucket marketing spend falls into. You don’t want to be too timid in not spending enough on subsidy initiatives and fall behind competitors while you might be able to simultaneously acquire users through these methods and improve the product. There is also a positive externality of this type of  marketing spend on the R&D department of a startup by giving engineers more touchpoints and data from new, early users to better understand how they might be able to improve the user interface and technical capabilities to better meet demand in particular markets. The below is a helpful framework to conceptualize some of the tradeoffs and tipping points that delineate customer acquisition and education costs for early product subsidies and discounts.

To the extent they use subsidies, companies should generally target to spend these sales & marketing dollars to acquire customers not to primarily educate customers broadly about their competitive market from which a peer could benefit. In the above, the switching factor refers to the probability that a customer will switch if the subsidizing company delivers on its promised value proposition. The product quality score indicates the probability of the company realizing its value proposition in a given customer interaction depending on the reliability and quality of its systems, technology and organization. These two factors are grossed up by the incremental probability that the new short-term users acquired from the subsidy will provide valuable insights to help the company improve its product quality score. This combined factor must be greater than your company’s churn rate, implying the probability of keeping the new customer is greater than the probability of losing it. It must also be greater than your competitor’s churn rates (assuming churn rates are a proxy for customer satisfaction). However, given there may be switching costs inherent to the product category, it is important to deflate even a superior competitor churn rate by a probability factor representing this switching cost deterrent that may be structural to the product category.

If one of these two conditions is not met, then you may be spending your marketing budget to subsidize broader industry education and evangelization. In that event, it is likely a better investment to relentlessly focus on execution to drive improvements in your product quality score to realize more of your promised value proposition. This product and operational investment may generate more benefit in the form of viral word of mouth marketing than a premature subsidy program. This is especially true for eye popping value proposition claims which have the potential to spread like wildfire on social media, such as the prospect of having groceries delivered in less than 30 minutes.

A Trojan Horse to the B2B Opportunity

Once a digital first, vertically integrated grocer establishes a strong network of food producers and manufacturers combined with a logistics-based moat, it would be quite logical to expand into the B2B segment. Currently the B2B segment in hospitality, restaurants, pharmacy, etc. is served by wholesale food distributors such as Aramark and US Foods. Especially in urban areas where they have hyper-local fulfillment networks built out, grocers like DIJA and Jüsto will be best positioned in the future to leverage their platforms to conveniently and cost effectively serve businesses, not just consumers. BigBasket, a leading online Indian grocery delivery business, successfully made this pivot by expanding its initially consumer offerings with bigbasket business, a wholesale vertical focused on serving restaurants and hotels. BigBasket made the shift by building new infrastructure on the business side, but new hyper-local consumer fulfillment hubs can be built with the foresight to be able to also handle business volumes. I also think there is an interesting opportunity to apply the group discount buying model in a B2B context where the volume throughput is even higher and group discounts can be bundled across B2B and B2C orders. Expansion into B2B not only diversifies revenue to a new vertical but helps to strengthen the grocer’s moat with food producers and suppliers by consolidating their wholesale and consumer food volumes for a region into a single source.

I am also excited about the opportunity for the integration between B2C and B2B food businesses to unlock greater efficiencies across the food ecosystem. For example, a large portion of food deemed no longer restaurant quality is thrown away by restaurant businesses but could actually be resold to consumers at discounts enabled by secondary markets powered by digital grocery brand portals and proprietary logistics networks.

The Importance of Personalization in System Design

Both of my discussions with Ricardo and Alberto got me thinking about the importance of localization and customization of system design especially for businesses with a substantial offline component. Jüsto’s proactive choice to locate its produce drop-offs by existing central markets substantially reduced friction for its growers by enabling them to simultaneously make deliveries to Jüsto and the central market and to keep with historical grower trade patterns and traditions. This made commercial selling economically viable for more growers and is an interesting case study into how smart network and system design can facilitate the expansion of a supply network by localizing network processes to particular markets. Relatedly, for partially offline businesses, there is an interesting opportunity to leverage AI and data to not only enhance the customer experience in a consumer-facing fashion, but more efficiently design offline networks and plan where to build subsequent network nodes. Density of user transaction data and supplier purchase order data can be highly predictive of which dark store nodes it makes sense to invest in building out next. These rich data sets ought to be used to drive smarter capital allocation decisions as the logistics network is built out to drive more ROI on new nodes. This helps reduce net effective capital intensity through network optimization. There won’t be one cookie cutter approach to choosing the next offline dark store node. Businesses should aim for their planning process to scale more efficiently as each node is added, but shouldn’t expect that same system will be as effective if copied / repeated in a new local market. When it comes to system design in bespoke markets, scalability does not equal repeatability

All Innovation Armory publications and the views and opinions expressed at, or through, this site belong solely to the blog owner and do not represent those of people, employers, institutions or organizations that the owner may or may not be associated with in a professional or personal capacity. All liability with respect to the actions taken or not taken based on the contents of this site are hereby expressly disclaimed. These publications are the blog owners’ personal opinions and are not meant to be relied upon as a basis for investment decisions.