Dark kitchens and the default path

Dark kitchens and the default path
“Imagine the most profitable restaurant in your city has no tables, no chairs, and no patrons. Chances are you just pictured the future of dining.”

That’s the first line of Do you know where your food comes from?, one of my favorite blog posts–back in April 2019. Things have changed quite a bit since then. And last week, Sifted published a comprehensive report on one of the hottest markets right now.

The future of on-demand food delivery.

Food delivery plays a sizeable role in European tech, with household names like Glovo, Wolt and Delivery Hero. It accounts for 4.6% of the total capital invested in Europe in 2019 – €1.6 billion out of the €34.3 billion.

The report outlines three predictions that they believe will play out over the coming years (and a bunch of puns):

  • Dark kitchens are going to be big
  • Brands are going to go at it alone
  • On-demand grocery is going to get interesting

The report is fantastic – European tech has been craving quality content for a while – but overly bullyish. I guess that’s what you get when you ask the Head of Asia Pacific, Europe, Middle East and Africa for Uber Eats about the future of his industry. I tend to disagree a bit, and think we’ll see some consolidation in the space.

But there’s one thing I think they got right – dark kitchens.

Welcome to the world of dark kitchens

Here’s how I described dark kitchens in the past:

"A cloud kitchen is a food venue that doesn't have a location for pickup or seated venue and are offered exclusively through food marketplaces."

Back then my point was a potential winner could be the infrastructure layer that powered the entire industry – the AWS for dark kitchens.

My point holds. But I think I was missing a bigger one: the combination of food delivery apps and dark kitchens can kill home cooking and change the way we consume food.

Food delivery apps can integrate vertically and control production on top of the already existing distribution by opening up dark kitchens and displacing competitors.

Old graph from ST19 // LINK

If handled correctly, this is a game changer for a bunch of reasons.

First, food delivery apps would become both sides of the marketplace, pushing down costs and increasing revenue.

  • as a food delivery app, they can retain all the revenue instead of sharing it with the restaurant (remember, they are the restaurant)
  • as a restaurant, you can produce food  at a much lower cost than a traditional restaurant because there’s no front-of-house personnel, rent is lower, insurance is more affordable, etc. Not only you retain all the revenue, but your COGS are significantly cheaper than a traditional restaurant AND you control the marketplace where this purchase happens.

This cost reduction can be absorbed by the business, or passed along to the consumer in the form of lower prices.

Second, food delivery apps can iterate at a ridiculous speed, and the speed of feedback cycles if one of the best levers you can pull when trying to improve a system.

Imagine an old hotel shower where water temperature takes a while to respond to your faucet turns. You’ll spend most of the shower turning the faucets left and right frantically, overshooting and undershooting to get the right temperature because it takes a freakin’ minute for the water to adjust to your inputs.

Same with feedback.

If you are a food business on a food delivery marketplace, there are two things to optimise for:

  • New orders. This depends largely on how your brand resonates with new consumers.
  • Repeat orders. This is affected by food quality. If people liked your food, they’ll order again.But let's talk about new orders.

If a regular restaurant with a front-of-house wants to try new branding and position to increase new orders, they need to hire an agency, change their physical space, and fire the chef to bring on a new one.

If Deliveroo wants to try new branding to increase new orders, they could ask their Product team to put a bunch of designs together, split test the most performing one on the app, and funnel all the orders through the same operation.

Example of different Ramen restaurant variations

Third, food delivery platforms have loads of data to inform decisions. They know what kind of food sells, at what price point, and at what time of day and there’s nothing stopping them from slowly building delivery-only, more affordable competitors to well-known restaurants.

In short, they can pull an Amazon. You can claim it’s anti-competitive. It might be. But for the consumer it just works. And it’s hard to make an anti-competitive claim when the consumer ends up better off.

When you top of all that with the fact that food delivery is getting faster every day, you have a lethal combination.

Look… this is not a walk in the park.

The real complexity here is operational: it’s a lot harder (and expensive) to scale physical facilities at venture speed than software. Plus, this also distracts from competing against each other.

From the report itself:

“It’s a challenging and fairly capital intensive model to roll out, and is taking longer than expected.” - Yannis Alivizatos, General manager for Deliveroo Editions

But it all comes down to defaults.

The default path for getting food

In an episode of Invest Like the Best, Matt Clifford from Entrepreneur First talks about the idea of default paths and how there is opportunity in flipping them around.

“I think the idea of ‘What is the default path?’ is the really important one. If the default path changes, then there’s an opportunity. The inspiration for why we started Entrepreneur First, Alice and I met at McKinsey. Why were we working at McKinsey? Default paths. I don’t even remember filling out the form. I graduated from Cambridge in 2007 and that’s what you did.”

He goes on to explain how Paris has a highly stratified, rankings-obsessed education systems where the prize for being amazing is to go to the next level until you become Emmanuel Macron.

The default paths for ambitious people in France is set. High-school, exams, grand école,  so on and so forth.

Side note: you can read more about this on the best book on a culture I’ve ever read, Au Contraire: Figuring out the French.

For Matt Clifford and Entrepreneur First, the interesting thing in talent investing is finding ambitious people whose default path is set and show them they can be founders.

For food delivery apps, there is a huge opportunity in showing people that the default path to obtain food is not to cook, but to order.

But a few things need to happen before we can get there. Let’s do a fun exercise and look at food sources across five different variables:

  • Quality of food. How good is it? How tasty is it?
  • Health score. How healthy is it? Where do ingredients come from?
  • Price. How cheap is it compared to alternatives?
  • Speed. Can I get it now? What needs to happen for me to get it?
  • Reliability. Can I make sure that I always get it when I wasnt it?Now let’s compare food delivery apps against cooking at home using a value curve.
Value curve of home cooking vs. food delivery now

In the near future, dark kitchens will help food delivery cross a threshold where ordering from them will be cheaper, better, faster, healthier and as reliable as cooking at home.

Here’s Oscar Pierre from Glovo, one of my favorite founders.

“The number one reason why people don’t order more food delivery is because it’s expensive. In Spain, you’re paying 12 per person, more or less, which is expensive. And the reason it’s expensive is because we’re still paying the prices of running a restaurant – which has nothing to do with delivery.”

There’s an old blog post by Bill Gurley on how to value Uber called How to Miss By a Mile: An Alternative Look at Uber’s Potential Market Size.

In that post, Gurley suggests that when you significantly improve an offering you generate new use cases and expand the market along the way.

In Uber’s case, people now send their kids to school on Uber. In food delivery, people stop cooking and order Glovo.