Finally, people are ‘getting’ groceries!

We at Kinnevik have been intrigued by the food space for some time, because:(1) the Total Addressable Market (TAM) is massive; food constitutes by far the largest share of the household’s somewhat discretionary spend, (2) the sector is Ripe for Disruption; it’s still dominated by outdated incumbents with massive fixed cost bases and legacy processes that hold back innovation and self-cannibalization, and (3) we see Strong Secular Tailwinds; sustainability, health and consumer preferences were already fuelling progress, but the pandemic is turbo-charging these forces. We believe this holds true across the broader food ecosystem, where the two big value buckets (highly simplified) are downstream (consumer facing) and upstream (production of actual food).

Back in 2017 when we started looking at the food space in a more deliberate manner, we as a firm were quite focused on backing consumer brands. As such, we started our journey into the wonderful world of food by digging into the downstream. This led to our initial investment in Oda (the dominant online grocer in Norway) in 2018, followed by our investment in Mathem (the leading and only pure-play online grocer in Sweden) in 2019. In retrospect, these investments may seem uncontroversial. But back then, the space was far from in vogue (perhaps as some investors were still licking their wounds from the Webvan adventure 20 years ago), and we consequently had loads of interesting conversations with various stakeholders.

In general, it seemed like people intuitively got that the grocery segment has some differentiated ecommerce characteristics like big basket size, high frequency and zero product returns. And to some extent people bought into the notion of ‘platform value’ derived from the proprietary last-mile and native app relationships with the customer. However, there were (and to an extent these persist today) some questions around (A) the viability of the business model, (B) the size of the addressable market, and (C) what the prospects of pureplay online grocers are in said market. Below I have outlined my reflections on some of the most frequent questions in these areas.

But isn’t grocery shopping a low margin business?

Not necessarily. Yes, offline incumbents tend to have razor thin bottom-line margins. But that is not really the relevant reference point for online players, since the cost stack below cost of goods sold (COGS) is quite different. As such, the relevant sector reference point is rather the product gross margin. The product margin clearly varies greatly by market and segment, but a ballpark figure of 30–40% is achievable in the large ‘supermarket segment’ in many regions. In addition, online players tend to have a considerably higher share of fresh produce and other higher margin stock-keeping units (SKUs). Add to that the scale effects of centralized larger Customer Fulfilment Centers (CFCs) (compared to the smaller brick-and-mortar stores), which on the COGS level include more efficient inbound logistics and considerably less spoilage. And in the also fairly sizeable ‘convenience store’ segment, product margins can climb even higher. As such, groceries are, from a product margin perspective, in apparel territory — a category that has arguably quite successfully transitioned to and scaled online.

But groceries are low average selling price (ASP) products — can home delivery really be profitable?

There are two main variable cost buckets between gross product margin and variable contribution margin before marketing cost (CM2), namely the ‘picking cost’ and the ‘distribution cost’.

Picking cost as a percentage of sales is largely a result of (i) picking efficiency, (ii) wage level, and (iii) ASP. These are all areas that can be optimized, but within a given market and context, picking efficiency is really where you can establish a competitive edge through technology and operational excellence.

The industry standard metric for picking efficiency is UPH, measuring how many Units are picked at a CFC Per Hour. Ocado (who popularized the term) had a UPH of 169 in 2020 and are targeting 200+ over time. Our portfolio company Oda recently announced that they are well beyond 200 already today! 😉

Wage levels clearly vary greatly by country. Even amongst affluent European countries, the delta can be more than 2x. Consequently, in high wage countries (like Norway), the impetus for optimizing the other two levers is arguably stronger. That being said, wage levels will be similar within any given geography, which is ultimately what matters from a competitive perspective, given this category requires local fulfilment.

ASP (Average Sales Price) of course depends on the assortment your customer segment is buying, but also how you bundle. If you bundle 6 apples in one pick, that helps your ASP but to some extent hurts your UPH, so it’s important to understand both metrics deeply in order to assess the true potential of any particular operation.

Since all three parameters vary across markets and across players in each market, picking costs for ‘full assortment’ players consequently vary greatly, but typically fit into the wide range of mid-single digit to mid-teens.

Distribution cost as a percentage of sales is largely a result of (i) deliveries per hour, (ii) wage level, and (iii) basket size.

Achievable deliveries per hour is obviously affected by the geographic density of your customer base and the size of your business, because ultimately route density matters. However, drive time between stops is not the only variable — there is also finding parking, getting into buildings, taking elevators or walking upstairs — making ‘time at customer’ an important lever.

Basket sizes for ‘full assortment’ players vary greatly depending on actual assortment and customer segment, but typically span the broad range USD 100–200. Select players have smaller baskets, which makes it trickier.

Again, several factors are at play here, but even at relatively modest scale and consequent route density, ‘full assortment’ players with a centralized CFC model tend to get to high single digit / low teens type percentage of sales cost levels. With density, and importantly larger baskets, there’s potential beyond that.

As such, and depending on the product gross margin starting point, top performing full assortment scheduled delivery players can achieve a CM2 in the high-teens percentage of sales ballpark, and over time beyond that. However, for those who are less successful in execution across the key drivers of the model, the margin potential drops significantly.

But isn’t grocery delivery capital intense?

Well, it’s all relative, I guess. Given the high velocity nature of the business, you don’t really need much warehouse space, but rather fairly moderate sized CFCs. And, as proven by Oda, it’s possible to design very low capex automation.

Furthermore, given the high turnover of stock, you get paid from consumers well before you have to pay your suppliers. As such, you have negative net working capital and consequently release cash when you grow your business.

But have any online grocers really reached profitability?

A few years ago, this was a fair question. Ocado had of course reached positive reported earnings levels, but that was on the back of years, if not decades, of heavy investments in tech and capex. Most other players in Europe and the US lacked the sufficient scale.

As such, you had to understand the unit economics stack and the consequent operating leverage inherent in the business model. Today, it’s clear that well performing online grocers can break even at a topline of a few hundred million USD. And that past the point of break even, earnings will of course grow much faster than your topline, as fixed overhead and marketing shrinks on a percentage basis.

But can you actually scale the customer base at sustainable economics?

As always, not everyone will succeed. But for those who deliver a great service to their customer base, I’d make the following observations.

Like most investors (presumably), we like to assess the scalability of these type businesses by borrowing logic from subscription businesses and looking at the relationship between customer lifetime value (CLV) and customer acquisition cost (CAC).

The CLV is a result of (i) customer retention, frequency of purchase and basket size, or put differently your revenue retention, and (ii) CM2.

What we’ve historically observed in full assortment scheduled delivery grocers is that customer retention in monthly cohorts typically falls fairly sharply over the first six months to then level off and stabilize somewhere in the 10–20% range. For some a bit lower and for some a bit higher. The absolute percentage level is in and of itself of course not very relevant — it needs to be taken in relation to the CAC to acquire those customers (if you can acquire them cheaply, a lower retention is acceptable). What is very important though, is that once it has stabilized, it typically remains flattish more or less indefinitely. Now you are at your actual customer base, so the net cost (CAC minus CM2 to date) of the customers who have churned can be viewed as part of CAC for this core base.

The true magic occurs when you layer on the evolution of purchase frequency and basket size, which both tend to follow a steady upward trajectory, thus producing a U-shaped revenue retention curve, i.e., with negative revenue churn in mature customers. As a result, your long-term revenue retention can approach 50%; of course, half of that is just the mathematical effect of users on average joining the service half way through the first month, but still.

As such, assuming you have a decent CM2, the CLV is massive (for consumer ecommerce) so you should keep acquiring customers at pretty punchy CAC levels.

But how much of the grocery market is really going online?

The traditional grocery assortment encompasses food and other household consumables. A lot of it include branded and/or commoditized low interest products, which are often bulky and relatively heavy. As such, getting these items delivered to your home is a great service. Also, the larger scale and higher throughput of CFCs enables a broader assortment, particularly in fresh products, compared to what is sustainable in much of the brick-and-mortar footprint. Importantly, the model has a structural advantage as it pertains to spoilage levels, partly due to the larger scale and throughput, and partly due to the data driven nature of the business providing benefits when it comes to forecasting demand, but also managing stock levels through targeted offerings, etc.

Consequently, our thesis is that a huge share of groceries will go online already in this decade. As noted above, the tricky part is to predict the speed of this online transition. It’s often said that people tend to overestimate the impact of technology in the short term, and underestimate the impact in the long term. To the extent that holds true, it’s good to be a long term investor without fund life restrictions (as we are here at Kinnevik 😉). When we started deploying capital in the space, there was less visibility on that. The covid-19 pandemic has clearly accelerated the digitization, but it remains to be seen by how much. Some consumers will go back to offline, but many have created a habit of getting their groceries delivered at home.

In addition, there is a meaningful opportunity beyond the traditional grocery assortment. Having a last-mile fulfilment platform and a high-engagement native app relationship with a large customer base provides an unfair advantage to compete for adjacent household consumable categories, in particular commoditized categories like pharmaceuticals, since the delivery is already paid for and the customer is already largely acquired.

But aren’t the offline incumbents going to take the bulk of the online market?

Sure, that may happen in some markets and the UK has been a good example to date. However, in many European markets, the offline incumbents have inherent structural challenges to heavily investing and competing for the online market, in particular the home delivery segment. First of all, it largely entails self-cannibalization of a more profitable business (since they have sunk cost in bricks-and-mortar footprint). Second, the first point is often exacerbated by franchisee models and the like. And thirdly, they may lack the mindset, culture and team to execute on the opportunity.

But how about click-and-collect in store, where incumbents have a structural advantage through their offline footprint? Fair point, this model has proliferated quickly, in particular through the pandemic. But even pre pandemic, it constituted the bulk of the online market in e.g., France. This model does make some sense in certain rural areas where home delivery economics are challenging. However, the click-and-collect model is inferior to home delivery, and should stagnate or decline as the home delivery service offering evolves. Firstly, the customer experience is in most cases less appealing than getting home delivery. Secondly, it’s challenging to achieve attractive picking cost levels when picking in offline stores that are not designed for this use case.

Although some offline incumbents will undoubtedly build successful home delivery businesses, many will experience negative operating leverage as same store sales decline due to online migration. As we’ve observed in other categories over the past decade (e.g. fashion), profitability quickly evaporates.

But won’t Amazon ultimately win this category as well?

Amazon is clearly a formidable competitor in any category, and it has publicly stated for years that it needs to make it in the groceries and apparel categories to meet its growth ambitions. Yet, across most of Europe, it’s not really there yet. I’m sure there are many reasons, but one is probably that these are somewhat emotional categories. You’ll for sure trust Amazon to efficiently bring you books or electronics just as much as the next ecommerce provider. However, do you really want Amazon to curate your closet or pick out fresh local produce? That said, the broader threat of players like Amazon coming into the space does highlight the need for online grocers to build a differentiated assortment and a trusted brand.

But how about other online grocers coming into your market?

Building a profitable online grocer is complicated and even if you do it well it requires scale to break even. This is somewhat prohibitive to brand new entrants, except as currently observed in the instant grocery space. Beyond that, we are clearly seeing full assortment players that have succeeded in one country moving into new markets, including our portfolio company Oda going into Finland and Germany, and Czech player Rohlik moving into Germany, etc. The benefit of that approach is clearly that you can leverage your overhead and tech stack across a bigger TAM. However, groceries is a local game, both in terms of sourcing and fulfilment, so it’s a significant undertaking to enter a new country. In Oda’s case, we believe that they have a structural advantage at this point in time through their proprietary best-in-class full stack fulfilment model, which gives them the right to compete internationally.

And more recently, we have been getting the question: But how about the instant grocers?

Instant groceries is clearly on the rise in Europe, both through takeout platforms like Deliveroo, Foodora and Wolt, and from vertically integrated pure plays like Gorillaz, Flink, Weezy and Getir. As already observed with GoPuff in the US, early signs seem to indicate that European consumers are also excited about the instant delivery proposition. And with all the capital going into the space, some meaningful businesses will be built over the next few years. However, it remains to be seen who will be left standing and what the consumer offerings will look like, after what is sure to be an interesting journey.

The impact on full assortment players is hard to prejudge. As of today, the services are largely targeting different segments of the market and are in some ways complimentary. The full assortment players core segment is large households, typically suburban with kids, whereas the core segment for instant players is a slightly younger urban crowd, including single households and DINKs (dual income, no kids).

So, what’s next?

In summary, we are convinced that great online grocers can build really big profitable businesses.

However, as the online penetration goes deeper and the market matures, the nature of the game will evolve, and likely include further fragmentation. We’re already seeing innovation along all three classic ecommerce drivers: Convenience (e.g., instant groceries), Price (e.g., surplus inventory discounters and price clubs) and Assortment (e.g., vertically integrated farm-to-table players and private label led direct-to-consumer (DTC) brands. To some extent these services are complimentary, but they will also compete and ultimately morph towards each other. Over time, certain aspects of these services will increasingly become table-stakes, accentuating the value of building a trusted brand backed up by a quality assortment, because ultimately more and more consumers will care about what they put into their bodies.

As illustrated by the velocity and size of funding rounds in the space over the past few months, it seems like many investors are coming around. We’ve deployed north of USD 200m into online grocers to date, and continue to think it’s early days for us and for the sector.