Is there any reason to believe that online grocery now is a better business than when Webvan failed? originally appeared on Quora: The best answer to any question.
Answer by Jonathan Brodsky, SVP of Chicken Soup for the Soul, former Director at 1-800-flowers, on Quora:
Grocery is a pretty awful business, in general. Whole Foods, which is one of the best-of-breed, operates with about a 33% gross profit margin and about a 3-4% net profit margin. Publix operates about a 6.3% net profit, and Kroeger and Fresh Market are around 2% each (data courtesy of The Modern Financial Data). Those are tough numbers.
It’s also one of those weird businesses where scale isn’t necessarily great. A lot of us are used to the software business, where, essentially, you build something and then you profit from it for a long period of time. In grocery, though, expansion means more infrastructure. A lot more infrastructure. Yuval Ariav gets this completely right.
The infrastructure problem isn’t what you might normally think of, though. It’s not warehouses with freezers and refrigeration. While that can be expensive, it’s not that hard to find an empty warehouse in most cities in the world that you can put this into, and that’s not really a re-curring expense. And all grocery chains – whether they’re online or brick-and-mortar – need distribution hubs.
The real problem is delivering that food to the customer. But, first, a little more on grocery store economics.
In a traditional grocery model (and most chains have nuances on this – I’m generalizing) – you’ve got a central distribution facility that services some number of your brick-and-mortar stores, and those stores service a certain radius of customers (e.g., customers within 5 miles of that store). You’re pretty much assured that you’re going to fill up your truck with product for a store, and it’s doable to make sure that there are people at the brick-and-mortar store to unload the product and put it on the shelves. They then have to solve the problem of making sure that customers go and buy all the food — especially the fresh stuff that goes bad — quickly. This is called ‘shrink’ at some of these companies, but we can just call it what it is: food spoiling.
The online grocery model is supposed to turn this food spoiling issue on its head. If you can just find out what people want far enough in advance – and it turns out that you really don’t need too much notice – you can just keep the fresh stuff people want in stock, and then you eliminate this spoilage issue. You combine this with not needing to have a store footprint, and you should have much higher profit margins.
CSA (community-supported agriculture) boxes go a little further – they don’t try to predict your wants. They tell you, “This is what the farmer produced, and so this is what you get.” For the most part, though, online grocers (outside of companies like Door to Door Organics) aren’t trying to do the CSA model. They’re trying to mimic how you shop in a store today.
Predicting how much milk people want turns out to be doable. It’s hard, but it’s doable. And, if you’re in a big enough city, you can get ‘emergency’ deliveries if you’ve oversold your online inventory in a day. You’ll pay a bit more, but you can get it to the end customer. Online grocers can win on ‘shrink.’
However, the people at the good major chains are also really, really good at predicting how much milk they’ll need in any given week, and so this isn’t necessarily as big of an advantage as you would need in order to fund your customer acquisition costs (I wrote a little more about this on this Quora post).
So the question becomes – can you have a much better profit margin just because you’ve eliminated the store footprint?
At massive, massive scale, where your trucks are all filled, yes, you can. Before that, you have a problem.
My experience in fresh delivery is in delivering flowers, which have a similar problem to most other grocery – you’ve got a product that spoils, you need to keep it refrigerated, and it’s a bad customer experience (although a necessary one) to leave it outside someone’s door if they’re not home. On the (very few) times I delivered flowers, we could do about 200 deliveries in a normal work day, including back-and-forth to the main facility to pick up the product.
I would guess that an online grocer in a dense urban environment could do about 500 deliveries per day (informed by this Quora question), per truck, assuming that they can make a bunch of drop-offs at one apartment building at a time. There aren’t many places in the US that are this dense, which is why you see FreshDirect based in NYC and not, say, Los Angeles, where there’s lots of sprawl.
The economics of delivering in a refrigerated truck are pretty simple – send the truck out on a loop, and send the truck out full. You don’t want the truck to drive 50 miles straight out to make one delivery – all you’ve done is burn gas, man-hours, and waste a truck.
I want to compare the fixed cost of actually getting the food to a customer, because that’s where an online grocer should have an advantage. After all, that’s the premise of huge companies like Amazon – no retail footprint enables them to lower prices with equal quality. I’m going to compare Fairway to a theoretical online grocer, because Fairway has high-quality product and, more importantly, massive stores that should put it at a significant cost disadvantage.
In this article from 2013, it says that the market ask for new space for a Fairway market was $250 / square foot for their 11,358 square foot store, plus $100 / square foot for the 40,894 feet above it. That works out to about $6.9 million in annual rent (retail rental rates are generally quoted on an annual basis). I’m going to bastardize Fairway’s public filings and assume that their sales per square foot are $1,000 (sometimes it’s higher, sometimes it’s lower, but this is on the low end… and, again, these numbers are old). That would mean that, in this store, they’re bringing in $52 million per year in revenue in this store.
According to the ATRI (2014 data), the average cost of running a truck is about $67 per hour. This is for all types of trucks, not just local deliveries, and I think it’s reasonable to assume that a refrigerated truck is going to cost a little more to run (you’ve got to power that cooling system, and you need to be able to keep the cooling running for all of your produce), and that stop-start driving typical in urban areas is going to burn more gas. So let’s up this to $70 / hour, and figure out how much money you’re saving or losing versus a store.
Back to that 500 deliveries per day number that I had up above. Looking through the Fairway filing, I see a $42 average sale per customer per visit (they said that they had 15.6 million customer transactions in 2013, and so I just divided their revenue by 15.6 million). I’m going to assume that an online delivery store can take over that entire $52 million in sales, and that each individual sale is a delivery. This is just a mental exercise to try and do an apples-to-apples comparison. This works out to 1.23 million individual sales for that location, or 1.23 million deliveries.
Doing some more (iffy) math, this works out to 2,460 total truck routes per year, assuming perfect truck routes. Assume that these are spread out evenly throughout the year, and you need 7 trucks (well, 6.73, but you can’t have 0.73 of a truck) running 24/7 to match the deliveries. Let’s assume that people don’t want their groceries delivered in the middle of the night – you’ve got 12 hours in which to do deliveries – so 14 total trucks.
14 trucks * 12 hours * $70 * 365 days = ~$4.3 million. If we assume that both companies have similar costs of loading the trucks, maintaining their hub warehouse, etc., the online grocery store is beating the entrenched companies in cost by $2 million. Hooray!
Except, wait. This ‘analysis’ (I hesitate to call it that – it’s really incomplete) assumes that all of your trucks are full, and all of your deliveries are geographically concentrated in a small area. But that doesn’t make any sense. While you’re certainly going to steal some of Fairway’s customers, you’re not going to steal them all. And you’re an online retailer, which means that anyone can order from you anywhere in your delivery area, and you’re going to take that order. You might have orders in the Bronx, in Brooklyn, in Staten Island. So you’re really, really unlikely to be running full truckloads everywhere. It’s more likely that trucks that can make 400 deliveries per day. Now you need 19 trucks running, and your running costs are almost $5.8 million. Uh-oh. Now your competitive advantage is slipping away. The break point for the same amount of revenue seems to be around 22-23 trucks, or an average of about 300 deliveries per truck, per day.
What’s the right number of deliveries per truck per day? I don’t know. I assume that FreshDirect does this really well, and they’re closer to 500 than 300.
But New York City is one of the only places in the US – if not the only place – where you can consistently get this kind of density. The further spread out your customer base is, the more trucks you’re going to need to service them in a timely manner. You’re also going to be competing against grocery stores that have lower fixed costs – it’s a heck of a lot less per square foot where I live, an hour north of NYC, than it is in the city.
That relative value difference on your fixed costs is what enables one of these companies to survive, and causes the other to die. This is because the company that has lower fixed costs can afford to charge customers less, and customers generally like spending less money when they think that the quality is the same.
The only way I know of to overcome this relative value difference – that is, to charge a customer more – is to deliver a consistently better value. In grocery, that’s either super-high-end product, or prepared food. You can see from going in to almost every grocery store that they all like prepared food as a margin enhancer. And, obviously, prepared food delivery (e.g., pizza delivery) works everywhere in the country (heck, one of the start-ups that I helped run did this, and now it’s part of GrubHub).
I made a lot of assumptions in this post. We could argue that gas prices make a big difference (they do), but they’re also subject to massive swings. We could argue about whether the data I got on the cost of trucking is right or not. I don’t know. That’s why I got a third-party source that I can point to. But I know that the overall conclusion from the analysis is correct – I’ve seen this type of problem over and over in perishable delivery, and there isn’t a good solution outside of making the hub in the hub-and-spoke system as close as possible to your end customer.
So, my very long-winded answer to your question is: no. There’s no reason to believe it’s a better business, at least at national scale, than it was a few years ago.