Are cloud (virtual) kitchens profitable?
Travis Kalanick invested $150M into CloudKitchens, UnitedKitchens raises $40M — everyone’s talking about virtual kitchens…but is this concept too good to be true? In this piece, we’ll explore the unit economics & determine the feasibility of this new restaurant concept
Perspective: the following analysis assumes the perspective of a virtual kitchen operator in a traditional restaurant real estate scenario — while the unit economics & assumptions vary by method, they hold true to a large extent. An exhaustive list of virtual kitchen models I researched are depicted in Figure 1.
ASSUMPTIONS
- Any calculations/data presented assumes costs & revenues specific to the Toronto market (though this analysis can be extrapolated to cities of similar size, i.e. NYC, LA)
- The virtual kitchen operator seeks to begin operations in large North American metropolitans (where both delivery demand and lease costs are high) — for our analysis, let’s assume we are opening in Toronto, CA
- The virtual kitchen operator (landlord) seeks to open facilities in highly dense core downtown (where according to UberEATS itself demand for food delivery is highest, but, so are real estate lease costs)
- Virtual restaurant operators (restaurants) can only sell food via delivery apps (UberEATS, Foodora, Skip, etc.)
TLDR:
In the value chain analysis illustrated in Figure 2, it is clear that restaurant operators (tenants) cannot be profitable, whereas, most of the profit is captured by operators (landlords). Analyzing the above concludes the following : 1) restaurant operators must operate multiple brand concepts in one kitchen to be profitable and 2) a landlord’s profitability is entirely dependent on the success of its tenants.
While operating multiple brand concepts using one kitchen (and realizing cost savings by scaling fixed labour & rent) is becoming popular, there are significant operational risks involved:
- driving sales for new brand concepts will be extremely difficult in the short-run: existing restaurants may have strong brand equity with their original concept, but, this does not hold true for new concepts;
Remember: building a brand is not easy… especially in hyper competitive industries like food
- quality control and operating at scale is a challenge even for the most seasoned chefs/cooks— and in tiny spaces (virtual kitchens) this challenge is further increased
- as restaurants scale the # of brand concepts, there is a risk of diminishing returns due to concepts cannabalizing sales
- most importantly — the sales channel is both a blessing & a curse — delivery apps enable lower operational cost virtual kitchen models at the expense of …
being in control of customer relationships, capturing customer data & changing delivery app rules/fee structures/dynamics
In summary, while the lease economics can be profitable for a landlord, the landlord cannot control the profit factors (revenue & costs) for its restaurants — & unless its restaurants execute perfectly, restaurants will fail causing high vacancy rates & failure of the entire business model.
Virtual Kitchen Economics… Landlord
Figure 3 calculates annual profit and investment payback period for a virtual kitchen operator (landlord). As discussed earlier, this analysis assumes average real estate costs in Toronto, Canada.
Few things to note:
- Kitchen construction costs: the average construction cost as per industry estimates is $300–450 per sq.ft — however we are assuming a landlord can achieve economies of scale when constructing multiple (in our case, 9) kitchens as the incremental cost of materials is small;
- Equipment costs: in our model we are assuming the landlord is providing a fully turn-key kitchen to restaurants — a 400 sq.ft kitchen with basic equipment (fryer, grill, etc.) would be $25,000
- This analysis is super conservative at the advantage of the restaurant most kitchen operators offer restaurants much less value — smaller sq.ft, less equipment, higher lease costs. In our analysis, we assume an industry-leading 400 sq.ft size kitchen & all equipment provided
Looks great doesn’t it?! 3 year investment payback $225K annual profit! But wait — the landlord can only pull these numbers IF their tenants are able to pay the $4K/mo (or $48K/yr) lease! Let’s take a look at a tenant’s P&L…
Virtual Restaurant Economics… Tenant
A sensitivity analysis provided in Figure 4 below outlines the profit levels for restaurant operating one brand concept. Assuming the ‘Most Likely Case’ we arrive at a negative annual profit of about $10K!
Diving deeper into the below analysis (cost assumptions based on industry standards) we can identify the following:
- Despite lower operating costs in a virtual kitchen (vs. traditional restaurants) — unless labour and rent are scaled across multiple brand concepts, there is no profit to be made
- Traditional restaurants can leverage various methods (advertising, promotions, events, etc.) to boost traffic & sales, but these actions have no effect when your only sales channel are delivery apps
- 400 sq.ft is the ideal size to operate one brand concept (i.e. shawarmas) but when adding additional brand concepts, it requires extreme attention to detail to maintain process and not sacrifice quality control
In summary — operating a virtual kitchen is not profitable even under favourable conditions. The act of perfecting each & every process, nailing down food production at scale in a tight space while ensuring quality control & being able to scale multiple brand concepts is not something easily accomplished!
You remember that age-old saying… most restaurants fail? In fact, on average 23% of U.S restaurants will fail within their first year
So is the virtual kitchen model doomed?
Yes and no.
YES — because the underlying profit drivers for a landlord are occupancy rates & lease revenues. Unfortunately, landlords have little control over both profit drivers, which can be negatively affected if the initial base of tenants do not succeed (think about it — would you open a restaurant in a facility where 90% of other restaurants failed?)
Oh and if you’re thinking… “well my downtown core doesn’t have ANY vacant restaurant spaces, what gives? How can restaurants not be profitable if there is a tight supply of restaurant real estate?”
Well — the fact that there is a shortage of traditional restaurant real estate (indeed the case in downtown Toronto) only reaffirms my belief that restaurants need a dine-in space to survive! The composition of a restaurant’s P&L (Figure 4) doesn’t change much between a virtual and physical restaurant… except dine-in, walk-in, event revenues!
and…
No — because the illusion of being able to open a restaurant without any upfront capital expenditure will always attract tenants. Unfortunately, this cannot be sustainable in the long-term. So where do we go from here?
We explore non-traditional methods of operating virtual kitchen business models that lower the infrastructure & lease costs for a landlord … and … thereby, decreasing operating costs for a restaurant. Recall Figure 1, non-traditional methods may include operating in parking lots, or in food trucks (dare I say shipping containers?) — and perhaps this may be the future of virtual kitchen business models.