Nick Sleep Internet Scale Laws
As we've been researching online grocery, this comment by Nick Sleep in 2007 on internet retailing comes to mind:
It seems to us that the basic building block of internet retailing, its skeletal structure, is far more robust, scalable and cheaper than the high street equivalent. In other words its power law is very high, and implies that businesses with the simplicity of operation as say, Amazon.com, have a shot at being far bigger, quicker and more profitable than their high street equivalents. - Nick Sleep, Nomad Investment Letters
Centralised fulfilment is a core advantage for pure-play internet retailers. Suppliers ship direct to centralised warehouses and then the retailer ships directly to customers rather than to retail stores. This eliminates the most expensive part of the supply chain.
Stores are also a complex operation that determine the quality of the customer experience. The staff’s behaviour, aesthetic design of the store, and available assortment creates a different experience for each customer. These variable factors are hard to manage consistently at scale.
The variables that define the online customer experience are largely fixed. The website UI/UX cost is fixed and the same experience for customers globally. The product range and delivery experience are determined by the scale and efficiency of the logistics infrastructure. As Sleep suggests, these fixed variables make the ‘skeletal structure’ of internet retail inherently more robust and scalable than traditional retailing.
But how does this translate when selling groceries online? Does a pure play grocery retailer have a chance to win as much share online as Amazon in non-food retail?
UK Dark Grocery Stores, Centralised Fulfilment, and Ocado
We interviewed an executive with 30 years experience at Sainsbury’s to understand how UK grocers approach online. This executive rolled out Sainsbury’s first dark store 20 years ago and has been heavily involved in Sainsbury's online grocery offering for 15 years.
The UK is a mature online grocery market. Sainsbury’s and Tesco have been operating online for over 20 years from both dark and traditional stores. In 2020, UK grocery ecommerce penetration was 7-8% and Sainsbury’s recently reported 17% of 2021 grocery sales were online with 14% delivery and 3% click-and-collect.
The top UK grocers do not operate a centralised fulfilment model. Tesco fulfils online orders from 7 dark stores and existing larger stores in the UK. Sainsbury’s utilises 250 of their 600 large supermarkets across the UK for online. Both Sainsbury’s and Tesco, roughly 45% combined market share in the UK, are prioritising shipping from existing stores closest to the customer rather than from a larger Ocado-like centralised warehouse.
If we continue to see the trend towards immediacy and same day delivery, being close to the customer is crucial. Existing retailers with large store footprints effectively have fulfilment nodes close to customers. Unlike in non-food categories, the store footprint could be more of a strategic asset rather than a burden. This logic has led Tesco and Sainsbury’s to spend 20 years optimising in-store picking to leverage their fixed store base rather than adopt a centralised fulfilment option like Ocado's OSP.
Ocado may not be a fit for scaled grocers but it's attractive for pure internet retailers or those with a small store footprint:
Ocado, as a standalone model, works really well when you want to put a brand down in a market where there isn’t already a physical presence, or you are a retailer that is last to market. I think they do an incredible job for Marks & Spencer and Morrisons, because they are the latecomers to the ecomm party and have needed to scale operations and I think Ocado is a brilliant solution to help those retailers get up and running. What is really smart about the Kroger Ocado move is that Kroger would have had a choice; put down a supermarket site in Florida or put down an automated facility that enables them to distribute their groceries to a far larger catchment area. Because they don’t have a physical presence in Florida, they’ve gone for the option of building a fulfilment center, to distribute from there.
Impulse Buying and Grocery Assortments
The online buying experience is also very different to walking the grocery aisles. This comment from our recent interview stood out:
If you think differently about online, it doesn’t necessarily have to be a complete replication of a supermarket. It is becoming a channel of its own nature. You can start to curate a range which is more finely tuned to what your customers really want. You don’t have to carry 30,000 SKUs because you are trying to cater for so many different preferences and choices and you’re trying to encourage customers to buy more. A broader range creates impulse. Online, that impulse disappears. On your checkout page, you might get a pop up to say you may have forgotten a few things, but the customer journey is less impulsive. Therefore, 30,000 SKUs is not necessarily that important anymore.
Historically, as non-food categories moved online, selection has widened. Wider selection drives higher conversion, larger basket sizes and higher customer LTV. However, the optimal pure-play online grocery model could be to rationalise the assortment online. One insight from our recent interview is that online ordering seems to reduce impulse purchasing. One-click repeat ordering, automatic refilling, and a visible price basket price all reduce impulse behaviour and order variance. This is the type of ordering behaviour Sainsbury’s saw online:
When you are properly meal planning and placing an order online, you are also seeing the shopping bill tally up alongside. In my experience, what we were seeing in Sainsbury’s was that customers were pretty much spending their budget, whatever that budget was. Whether that was £120 or £60, there wasn’t a great deal of variance, week to week. They were clearly managing to a budget and filling their basket accordingly.
The CEO of Oda, the Norwegian online grocery, recently reported that ~400 SKU’s account for 95% of the average Norwegian online grocery basket. If you were building a pure-play grocery retailer, it seems optimal to reduce the SKU count closer to Aldi and Lidl’s 2,000 products rather than Tesco's 100k. It’s also easier to pick and ship fewer perishable items to customers. This is why we like Goodfood’s strategy of centralised fulfilment, ~4k SKU’s, and 100% private label. The higher private label margins reduces the pressure to hit Ocado like automated picking efficiency. The challenge for Goodfood becomes a merchandising problem more so than a picking challenge.