CoStar's London office at The Shard was the venue recently for a wide-ranging debate with experts on how grocery retailers are using physical stores and online orders to attract customers, with an eye firmly on what landlords and investors need to do as the likes of Amazon and Tesco rapidly change their business models. US expert Bill Bishop joined the debate via video link from CoStar's offices in Chicago to update on on how the market is developing on the other side of the pond and what lessons can be learned from both experiences.
Panellists - Ben Green, Steven Noble and Steve Windsor, partners of Atrato Capital, the investment adviser to Supermarket Income REIT Plc, Will Treasure, Director, Operations, Javelin Group, Bill Bishop, co-founder of the US’s Brick Meets Click; Moderator CoStar News editor Paul Norman
PN - What is omnichannel grocery and what is the key challenge for real estate?
Steven Noble –I would define omnichannel grocery as the integration of online with physical store sales. Those two can be viewed as competing against each other. However, in some sectors they can integrate and complement one another. My own view is that omnichannel is not a sales methodology but more a business model which is fast becoming an investment class in its own right. And what we see uniquely in the grocery sector is just how well those two can integrate.
Large grocery stores are increasingly being used as last mile fulfilment centres for online orders, which are fulfilled in the store and then delivered direct to the home or made available to pick up through a click-and-collect service. And we see that in the UK it was Tesco that pioneered this model. Now, across the world we see that model being copied, with major grocery retailers utilising their property footprint to build a network of omnichannel stores and, hence, we see this emerging as a new property asset class.
Bill Bishop – I think of omnichannel grocery as the integration of physical and online retailing with a primary focus on the store. It’s the first generation in the evolution driven by the digitalisation of retail. It is understandable that the store has played a central role up to this point and that will continue. Over time, though, I think there will be a number of changes that will impact the omnichannel value proposition.
One thing that needs to happen is a reduction in the excess real estate capacity that we have in the US. The stores are also going to have to become more welcoming and provide good reasons for people to visit them when they have the option to shop online. Over time it is likely that the physical and online retailing that is omnichannel grocery today will become part of an even larger integrated system, namely a customer-focused system that serves a greater share of household needs.
The key challenge omnichannel grocery poses for real estate is providing the flexibility needed by successful retailers as they make the changes required to keep up with rising customer expectations, such as changes in store size and integrating the store to support the personal supply chain.
Will Treasure - In terms of the UK market the share is 8% in terms of online grocery. This is about the biggest in the world, with maybe China and South Korea a little bit ahead, but certainly in terms of the Western world it leads the way. This year, something in the region of £13bn to £14bn of groceries will be ordered online for delivery or collection. Historically, for quite a number of years the market was growing at 15% a year and, actually, in the last couple of years is it has slowed down to around 8 to 10%, because there are major supermarkets, particularly Tesco, that decided to prioritise profitable growth rather than non-profitable growth and that had quite an impact. So, you could see it continuing to grow at 5 to 10% per year. It is hard to guess how long that is. I think 16% of the market is pretty realistic but there will come a point at which it stops growing. Where that is though I am not sure and we are quite a long way from that.
The UK has quite a few advantages. It is a very concentrated market, with a very dense population and very high grocery margins. And a dense market means that delivery is relatively cheaper than elsewhere. In an affluent country, Tesco, the number one grocer, decided to get into it back in 1998. The US really woke up in the last few years and Amazon buying Whole Foods really was a real big shock to the system. It is growing very rapidly. Will it catch up with the UK? I don’t know because the economics outside the big cities don’t quite stack up – but it is certainly growing very quickly.
PN - What is the breakdown of click and collect and online?
Will – 10% to 15% for the UK.
Ben Green – How does that contrast to the US?
Bill – The penetration of online grocery in the US today is about 5% but there has been a big push recently and that number is expected to increase to over 8% over the next three years.
Now, most of the online grocery orders - about 75% - are delivered, with this work being done mainly by third-party services like InstaCart and Shipt. The challenge with delivery is that it is expensive to execute and using third-party suppliers means that the retailer loses control of the customer experience.
We’re now seeing rapid growth of pick-up through click and collect, which costs less to execute and the savings are being passed along to consumers. For many, this provides a real incentive to shop online and pickup.
Ben Green – The opportunity is to identify the stores that are genuinely omnichannel; that is something the real estate market hasn’t really caught onto yet. We believe those have a level of future proofing that non-omnichannel stores don’t have.
Steven – An example of that is our store in Ashford - 25% of sales is online, we believe, which is a major revenue driver for that store.
Steve – In the property market there is no differentiation of this.
Bill – The fundamentals of good commercial locations, such as the population density of the trade area, the accessibility, the degree of competition, are, I think, amplified in an omnichannel environment. I don’t believe this story has sunk in yet, because there is so much preoccupation with online shopping. Until the economic geography is integrated with the online economics, the value of superior real estate will not be accurately reflected in its pricing.
Ben – In the UK, 74% of all online grocery is fulfilled from a store.
Steven – But that increases to approximately 90% for the big four operators, with Ocado the difference.
Ben – So it is completely different to non-food.
Will – If you take out Ocado and you look at the major supermarkets, the last calculation we did it was that about 85% comes from stores and 15% from dark stores and dedicated warehouses. The economics are, if you have an established space of supermarkets, it is cheaper to pick and deliver from those supermarkets than it is to set up a dedicated warehouse. That pretty much holds true everywhere and that has been confirmed by Amazon’s behaviour since it bought Whole Foods in the US. What it has done is roll back some of the Amazon Fresh in the cities and put it in the stores. It plans to expand Whole Foods into delivery and pick up from all the Whole Food stores which reinforces that.
PN – Is online profitable?
Will – No. The numbers vary from net net of plus 2% which is the best we have seen down to about -10%. It is very hard to make money from it because you are picking 40 to 45 items – UK numbers around £1.90 to £2.00 per item. You store those items in a three-temperature regime – ambient, chilled and frozen - and you deliver them in a three-temperature regime. That is just very expensive to do that. They are very low value items and they are quite difficult and heavy to handle. Grocery online is primarily a logistics business and, if you look at the P&L of an online grocer, two thirds of the costs are around the picking and delivery, and one third is the ecommerce website. I think there are number of things you can do to make online grocery profitable.
There are nine different levers you can operate but it is very hard. Inherently, it is expensive by its own nature. Click and collect is cheaper and the reason why it has been more popular in the US and France is because it is cheaper - instead of spending £8 or £9 or £10 on your delivery, you maybe spend £1 to £2 with a customer collection and, even though you have no delivery fee, the net net is maybe £3 per order better for collection. The economics work better for collection, customers love delivery. And because we are now all offering delivery here it is very hard for the retailers.
Interestingly, if you are getting a 10% negative net margin and it is 1% of your sales you can happily ignore that. If it is 10% of your sales - and it is more than 10% in Tesco - you can’t ignore it, as it is bringing down the profitability of everything. And that is why the supermarkets are focused a little bit more on making profit than growth over the last couple of years, led by Tesco.
So what does that mean? They have raised the delivery fee, they have taken away some of the offers around the subscription model and reduced the attractiveness of those. Now differential pricing is the big question mark for all the grocers. Knowing that online grocery is inherently more expensive, knowing that the customers won’t pay the high delivery fee - maybe £6 is the most they will pay - the only way you can really recoup the gross margin is by looking at differential pricing online rather than in stores. Now, that is really hard because your customer may shop online on Thursday and may go in store at the weekend to pick up some extras. It is a hot topic. We have seen it work in Australia where it has been done successfully and, in fact, Coles have done that. My view is the jury is out but, undoubtedly, it is in on the radar of big grocers.
Steven – We have an element of that now with the differences between pricing in a supermarket and pricing in a convenience store. There is quite a spread between those two, as the grocer is building in the higher operating cost associated with the convenience store. But the grocers tell us the price elasticity online is quite tight in that they change the price slightly and they can see a big change in demand.
Will – Delivery fee yes; interesting about the price point. Will customers accept it? Interestingly, in the US it is a bit different. Aldi has gone into business with Instacart. Instacart has put up the website, Instacart will pick up the items, Instacart will deliver them for you. There are variants of that where the Aldi team does the picking and Instacart will do the delivering. Instacart uses a gig-based model whereby it is relatively low cost for doing the picking up and delivery. Now, Aldi and Instacart have gone into business in Chicago where Instacart puts a premium on the Aldi prices in a place where an Aldi customer is perhaps more price sensitive than most. The trial has been very successful. We will wait to see how this price sensitivity will pay out.
Bill – The differential pricing is in fairly broad use today. Our market is far less concentrated. But Instacart is a differential pricer and a number of regional players have done a good job with this. I think that the pricing is just one of the angles where we can make this business more profitable. The other angle is the logistical possibilities where we can increase the efficiencies, which can improve the profitability of the two-thirds that we described. That, in my opinion, has a high likelihood of happening as we introduce more automation into the process.
We have several players in the US working diligently to come up with alternative workarounds that improve the productivity and efficiency, so, if you guys are selecting 45 to 50 items, it will go to 60 items per hour in the US and then your selection but on the automation side it is 600 to 800 items per hour. So you have a capital trade off but you have tremendous increase in selection. But I don’t know whether you want to go now or later. To me, looking at automation and various approaches, is it is likely to be the way in the US.
I believe Amazon will build bigger stores and look for bigger properties and have the capacity to absorb automation under the roof. That is the place to look right now. The other thing we haven’t factored in is there is quite a shift in market share. You have experienced a little bit of it but it has filtered down as the discounters have lost their steam in the UK. In the US the hard discounters are dramatically growing and we have other new players, so there is a move to subsidise profitability to promote your growth.
Steven – For a real estate investor, the more online comes instore, the more these supermarkets can begin to differentiate themselves as an asset class in their own right. The old rules around trading densities, rent-to-turnover ratios fall away because the omnichannel supermarket looks great on those measures relative to those that are still focused on the amount of footfall sales that come through the store.
Steve Windsor – And, Bill, do you think that as automation grows they stick with the omnichannel models or you get more dark stores?
Bill – I think technology will be deployed in the actual store. There will be some dark stores in the US but they’ll play a small role. There’s just too much economic advantage in fulfilling orders and retailing from the same physical unit.
Will – There are three suppliers at the moment. Alert Corporation, which is in partnership with Walmart, Takeoff Techologies and Commonsense Robotics - two US, one Israeli. I think it is quite early days - they are all in proof of concept. It looks like there will be significant cost reductions in that. But I think we need to see it in action and we need to see it really working. The key thing is, if it does work, it gives you quite high volumes in a relatively small footprint.
Coming back to how you think about the omnichannel store, geographic location is absolutely key to get the delivery density from there. Also space for your automation. You may not need a lot - 10 to 30,000 sq ft – as well as yard space, which is the bit people miss out. A general merchandise example is, if you have a typical retail warehouse supplying stores, it is 70% building footprint, 30% yard footprint. When you come to an ecommerce warehouse you really want it to be 40% building, 60% yard space – the yard space is needed for the vans. Because instead of having a relatively small number of articulated trucks coming in and out, you have a very large number of small vans coming in and out very frequently and therefore you need a lot of space. And the same applies in your omnichannel stores. If you have a great store with no yard space, you are going to really struggle to turn that into a high-volume, omnichannel store. So there are a number of different ways of thinking about the omnichannel store. And that applies with automation.
Ben I was going to say in a UK context because delivery is so dominant it amplifies the geographic advantage of the store, if it is costing you £9 or £10 for every delivery. It does not matter how much you save on robotics, your cheapest option is still the customer buying in the store if they don’t want to click and collect. At some point that price differentiation should be made clear to the customer.
Will – Yes competition and customer behaviour are what are driving online grocery. I don’t think that will change.
Ben – They are willing to go to Aldi to get the price but I wonder if they would be willing if you had differential pricing to go to the store to get a cheaper offer representing the true economic price. It will come down to how much they value their own time. If it is costing them more than one hours’ wages to get their delivery to them, then they will pick it up themselves on the way home from work. It is crazy to think I am going to have to work an extra hour just to get my groceries delivered.
Will – I think for online grocery your customers are, typically, families and, primarily, affluent families, as taking young children to the supermarket is not a nice experience, so there is perhaps an added incentive to have the delivery done. What you are saying economically makes sense and it is quite hard to see where we may end up in 10 years’ time.,
Steven – What informs the level of capital investment from the grocers will be how big online becomes as a factor of overall grocery. Right now, when we talk to the grocers, the level of capex they are willing to put into online is limited because the market share still isn’t big enough to achieve acceptable payback.
But how far do we think that needs to go before they start releasing capital expenditure into those technologies? IGD downgraded their forecasts for online growth. It seems like online is struggling to penetrate the same way it has in retail.
PN - Why has it had such a dramatic impact on retail and not grocery?
Will – In grocery in the last few years there have been three areas that have been growing – discount, convenience and online. Core, supermarket, shopper and supermarket sales are in decline and, as a result, the grocery market is overspaced for that core normal shop, hence the omnichannel. If you are a grocer and you have limited capital, you think how much do you apply online and how much do you apply in convenience and what can you do about discount. Not a lot about discount. You are looking at online and you are looking at convenience as your two growth channels.
Now, clearly with convenience, you hit a point where you can’t put down more stores. So what are you going to do with online? The problem with online is it is hard to justify a big capital expenditure in a loss-making channel. Just everything you look at is putting more money into something that is losing money. That is why there has been that resistance to putting capital into it. It is hard to predict the future. Because it is overspaced I think there will be a shake out of grocery. At what point and why are really open questions. But I am sure there will be. If then we get the right space in the UK for the sales, does that free up more capital for online? This the open question.
Steve – So what are they investing in? What are the key bits of automation? Is it Ocado and the picking and robotics and the human hand, or is it the conveyor belts and back of house racking, so all the baskets hit the van in right order?
Will – The micro-fulfilment automation that I have been talking about – those three technologies - is a three-dimensional grid. It is bots pulling crates in and out of that three dimensional grid and pulling them to the picker, so the product comes to the picker; the picker than takes the product, usually still by hand, and puts it in a crate which is a customer order and then it goes off. That is it. They are all run on that basic principle. Each one will say that their technology has particular advances that makes it unique.
Steve – And it is not that different to the Ocado bots you see on a YouTube video where they are flying around the 3D grid.
Will – The Ocado challenge is to ensure its customer fulfilment centres (CFCs) are built to process 100,000 orders a week. That gives some real benefits in terms of efficiencies but it results in high stem mileages for delivery. In other words, you are delivering up to 100 miles for your CFC. Your delivery costs are higher, although your picking costs are lower and highly efficient. That is the trade off with the Ocado product.
As online grocery grows, 100,000 orders might have to go 100 miles today, but in five or 10 years’ time you might only have to go 50 miles to do that same volume. Therefore, as density grows, it favours that Ocado model. But it needs a mature fast-growing scale market. Which is kind of where the UK is now but not where everybody else is. Those supermarkets partnering with Ocado, be they ICA or Sobeys or Casino, are essentially making a long-term bet that this centralized, high-volume fulfilment operation will be the right long-term solution.
Steven – Do you have a feel for what that market share would be to release that capex? Because, as Ben said, if you end up with differential pricings in store, convenience and online and, if online is at a price premium to recover the capex investment, assuming the 10-year payback for the technology, then there might just not be the take up because everyone will say it is easier to jump in the car and do click and collect. I think at the moment IDG is saying grocery in five years will still be less than 10% of the market.
Will – What I have seen is that, typically, clients have underestimated the growth of online which means logistically the pain they have had is because of that. So, who knows where online grocery will get to but don’t underestimate it. These technologies are called micro-fulfilment because they are supposed to be deployable at small scale. The principle is they are supposed to be reproducible and you can put them in 100 stores, design once and install 100 times. If that really works, then the economics can get significantly better – now that is a long way off.
Steve – And can they work alongside shoppers then or do you need a dark space?
Will – This works in your bigger stores where you are over-spaced at the moment. You want to take space that is not very productive, or, in fact, not usually missed if you take it off and use it, so it will be a separate area. You can think a bit further forward. John Lert, who runs Alert Corporation, in partnership with Walmart, has come up with a further concept which is that automated back-of-store has all the ambient slow moving items - what the Americans call centre store products - in there, and you can order online for collection in store, order online for delivery, arrive at the store and just put your order in and the front of the store is the fresh and interesting products. There are various versions of this and in China there is Hema. US and UK grocers are not exactly embracing this but 10 to 15 years ahead I would be thinking about what we can do with this.
Ben – It is interesting that the Ocado bet on densities relies on distance not getting dramatically more expensive which is not a given, when one of the big criticisms of online is it uses the public end goods and commons and doesn’t pay for it.
Will – Interestingly, there are very few environmental studies about online grocery. I have only seen one and that was about 15 years ago. The one I saw suggested that the van delivering 10-15 orders on a round trip has a significantly lower carbon footprint for delivery and less miles than 10 or 15 people customers driving to store. There is a strong argument that, although highly labour intensive, from a road use and carbon footprint point of view, it is relatively efficient. I think there is a sustainability argument. But the point about cost of delivery is interesting because, if you had asked me two years ago, I would say driverless vehicles would be here in 2020, but I don’t believe that now. I believe 2025 or maybe 2030. But when driverless vehicles come, the cost of delivery will fall quite substantially. That starts to improve the economics of the dedicated fulfilment centre.
Steven – What are your thoughts on the Ocado tie up with Kroger? It hit the headlines in the UK and no one knows if it is three CFCs or 20?
Bill – Well I don’t think we understand it either. The notion of this distance continuum is one way to explain it but I am not sure it is very persuasive. Kroger is a public company, so they have to be producing something with a higher degree of certainty. It is also one of the better run of the larger supermarkets in the US, perhaps the best run. I think Ocado will be used to modify its core delivery systems both to the consumer and to the stores and, if it can do that, it can dramatically change sales per square foot and inventory returns to metrics that we would use and focus on.
I suspect that sooner or later it is not going to be a long-term bet, more a short-term opportunity to use technology to re-engineer the details’ business model and inventory levels in the stores themselves. I don’t know how else it can work. We watched Ocado try to find a North American partner for five or six years and, with the exception of FreshDirect in New York where you have the population densities, there was no place. Sobeys, in Canada, which operates in thinly-populated markets, also seems like a stretch to me, but I have been surprised before.
So far, we have discussed automation primarily as a fulfilment process or a storage and retrieval system for fulfliling online orders, and I think that is 90% of it. But, if you take a look, for example, at what Walmart is doing, it is actually looking at the storage and retrieval as systems to expand the capacity of physical units to do online business. Ironically, it is staying in a manual select mode but expanding its storage and customer service with automation retrieval. That takes a little bit of space, so let us not forget the fact that automation can play a secondary role and, if you don’t want to sacrifice selling area, you could use it to drive sales on a per-unit-basis in each of those stores and increase the capacity; right now, Walmart probably has more stores that are maxed-out than most of us would realise.
Steve – Do you have a view on the number of these distribution centres that Ocado and Kroger can lead to? Ocado have been quite punchy saying around 20, while Kroger have been more conservative suggesting single digits.
Ben – From over here we envisage it works around New York and San Francisco with high enough population density but we cannot identify very many of those.
Bill – I think we are on the same critique. I don’t yet see how the whole thing bolts together.
Will – There is another level of complexity. Ocado has moved from running a UK business optimised for the UK to providing front-end ecommerce sites and back-end fulfilment for up to four of five different customers. What Ocado is facing is that everything that is optimixed for the UK is not necessarily right for other countries. For instance, in parts of Europe delivering in plastic bags is a no no. But the Ocado delivery is optimised via delivering in plastic bags in crates. That is one simple example of how each country will need something a bit different. That change is a really big change to be a supplier of services to running your own business. I have worked with a number of different retailers and home shopping businesses and it is really hard. The likelihood is that rollout will be a little bit slower than might have been expected as modifications have to be made before it can be put in to the US in Kroger’s case.
PN – What will be the impact on supermarket property?
Steve – More by luck than judgement, they built these big edge-of-town stores to optimise for 15 minute drive times. A few years ago they were too big as people no longer bought white goods from those stores. But, fortuitously and maybe with good planning all along, they have had a rebirth and they are actually even more valuable because you run that drive time equation on the way out just in reverse for the vans coming out. Over the course of time omnichannel stores mean they will need more square footage to put more automation in back-of-house, so actually they become a prime piece of real estate. From our perspective there is not a lot of understanding of this concept in the property industry. Most people think the larger out-of-town store is a dying thing which is definitely not true. There is no differentiation in property markets of an omnichannel store versus a non-omnichannel store. What will happen is that over time, as this feeds into the wider market, there will be a major differential and a new asset class for omnichannel stores.
Steven – What is clear is the amount of investment that must go into the site from the operators if online grocery grows, because the micro-fulfilment is a technology play and that doesn’t come cheap. Occupiers are going to be putting quite a lot of capex into these sites in the future to make them viable and from our point of view that is fantastic. It is not a five-year refurbishment cycle, it is a 10 to 15-year capital investment play.
Will – In grocery the omnichannel store is here to stay.
Steven - Which I think will, ultimately, back solve itself into rent. When you think about how rents are set, it is usually based on how much an occupier thinks they can turnover on the store. These sites will just increase their sales density and, assuming a constant sales-to-turnover ratio, rents will increase, and that is what is so interesting for investors
Steve – Ultimately, rents should be a function of store profitability. Now, with all this automation and the uptick of online, that piece of real estate should become more profitable to the operator, so they should be prepared to pay more for that piece of real estate. That will be slow in coming and rent reviews will not immediately do that binary shift, but you can see that trend.
Ben – We have an advantage against the US. Because we are a crowded island it is very difficult to move these things and most of the large sites have gone in desirable locations. It is a challenge that discounters are having, sourcing enough good sites to drive their expansion because Sainsbury’s and Tesco and Asda have taken most of the good sites already.
Bill – The fact that better properties will attract more capital investment bodes well for the stores that grade out well in the omnichannel environment.
But it is possible that over time some companies like Picnic in the Netherlands, and perhaps Amazon, will develop highly-efficient grocery delivery systems where the product never goes to the stores, which would have a huge negative impact.
Steven – What are your thoughts on Amazon and what is happening in the grocery space right now?
Bill – I really think Amazon has a long-term plan for grocery, and that its stepping stones near term for growth are omnichannel food stores. Eventually, it will want to own the complete distribution system and use its strength in grocery to give it the route density and delivery frequency that allows it to profitably deliver a range of other products to the home. There’s a lot of distance between the “cup and the lip” here, but it could happen sometime in the future. People would be wise to stay alert to signs of something like this taking shape.
Will – There is a bit of a holy grail, which is to say drop density is the key to low-cost delivery. If you think of Amazon, it has set up its own delivery network. It has Amazon Flex for the gig economy and various other things. Most of the Amazon parcels are now delivered by Amazon drivers in the UK. The end game is where you have grocery and general merchandise on the same van and you maximize your drop density by doing it together, and I do think there is a big prize for doing that, if you overlay on the subscription model. It is taking longer to get there then it would have liked. And the way it is going with Whole Foods suggests that grocery will be coming from the Whole Food stores and non-food from the warehouses which does not allow them to be brought together.