Are third-party delivery platforms really evil?
An analysis of own delivery, third party delivery, different delivery platforms, and unit economics
Highlights
- Online Food Delivery with massive growth, accelerated by Covid
- Third-party delivery platforms vs. white label solutions for own delivery — what options do restaurants have?
- Three dominant platform models
- Unit economics deep dive — the good, the bad, and the ugly
The past year has been extremely tough for the restaurant industry and I am very grateful that restaurants are slowly opening up in many parts of the world. Pickup & deliveries have been the lifeline for many restaurants and even if dine-in will come back very soon, I believe that pickup & delivery is here to stay. In the past year, there have been heated discussions around the business practices and fees of third-party delivery companies like UberEats and their effect on the restaurant industry.
In this article I will analyze the developments of the last year, I will shed light on the different options that restaurants have when it comes to food delivery, I will deep dive into the unit economics of restaurants and third-party delivery platforms and I will look into why restaurants often hate third party delivery platforms.
Disclaimer: I am the founder of a restaurant discovery platform called DiscoEat. We introduce yield management to the restaurant industry to help our restaurant partners to increase their utilization in off-peak hours. Covid hit us just as hard as our restaurant partners as we were a pure reservations platform until Q3 2020. We developed and launched a pickup and delivery platform in Q3 2021. We are NOT a third-party delivery platform like Uber Eats, Deliveroo, Wolt, etc. We send the orders of our customers to our restaurant partners. My co-founder and I were Managing Directors and Operations Directors of Foodpanda, which is why we know the business extremely well. I do not have any personal interest in either supporting or in bashing third-party delivery companies. You will find arguments in their favor and you will find arguments against these platforms. I am trying to be as neutral in my argumentation as I possibly can.
Online Food Delivery with massive growth, accelerated by Covid
No doubt, Online Food Delivery has been one of the winners of the Covid crisis when it comes to revenue growth. The food delivery market is expected to grow from €22bn in 2019 to €32bn in 2024 in the US and from €16.5bn (2019) to €28.6bn (2024) in Europe.
Food ordering is on the rise. Based on a market study in Germany, more than 53% of participants had ordered at a restaurant since Covid (vs. 40% before). Interestingly, most of the orders were still done on the phone (46%). Only 22% of participants had ordered from a delivery platform and only 17% ordered directly online from a restaurant website or app. Living in Berlin’s “tech bubble” — these numbers surprised me a lot. Only 1 in 5 people in Germany have used a food delivery app. While the absolute numbers may be different in other markets, the directional change too much more food ordering is the same across the world.
External delivery platforms vs. white label solutions — what options do restaurants have
Restaurant owners can build their own ordering service and they can work with delivery platforms. Of course, they can also do both. When building out my own ordering service I must keep in mind why people are using delivery apps:
- Large choice of restaurants
- Simple ordering process/ smooth UI/UX
- Simple payment, saving of payment details
A restaurant cannot mimic the first point, but it must have a system that is on par with the food delivery platforms on the second two points. In general, people love the convenience and are lazy and if they find it too complicated to order from a restaurant directly, they will use third-party platforms.
In order for restaurateurs to start their own ordering service, they first need to decide if they want their customers to be able to only order pickup or if they also want to start their own delivery operations. Whatever the decision is, in order for a restaurant to receive orders from customers, they need a software solution for their website. There are a ton of different white label delivery solutions in the market and what solution is best really depends on the needs of each restaurant. White label solutions differ vastly in terms of features and before deciding on a tool, restaurant owners need to decide what is important for them, e.g.:
- Type of Services: pick-up, delivery, both
- Web shop functionalities (e.g., multi-shop options, toppings & removables etc.)
- POS integrations
- Rider management tools, Rider Apps
- ERP functionalities
- Statistics and dashboards
- Integrations (e.g., with external delivery platforms, accounting integrations, etc.)
- Etc.
If a restaurant decides to (also) work with third-party delivery companies, they will have more or fewer choices depending on the city and country they are based in. In Germany, for example, restaurants in most cities only have the chance to work with Lieferando (owned by Just Eat/ Takeaway), who consolidated the market in 2018. The more platforms exist in a given market, the higher the competition between the platforms, the better the situation for restaurants and users as platforms have less power to charge high fees and fight for customers with vouchers.
There are three dominant platform types (Source). They differ in the depth of value creation on part of the platforms. The food delivery supply chain can roughly be categorized by the following steps:
- Menu Creation & Food Preparation — deciding what items will be on the menu, the price of these items, and the preparation of the food
- Marketing — attracting customers to place an order
- Ordering & Payment — the process of placing an order and paying for it
- Delivery — getting the food to the customer
A platform can be active in only one type or it can be active in several different business models.
Type I — pure platform model
These platforms are the first delivery platforms that have entered the scene around 15–20 years ago, depending on your country. They are the connector or lead generator between the restaurants and the users. The platforms attract users due to the vast choice of different restaurants and cuisines and provide a convenient ordering and payment process. The restaurants do the actual deliveries. Examples of players are Seamless & Grubhub in the US or Lieferando and DiscoEat in Germany. Those platforms usually charge between 10–13% commission (in Germany) and a payment fee to the restaurant.
Type II a — “Third-party delivery” business model
In the last 10 years we have seen the rise of third-party delivery platforms like Uber Eats and Deliveroo. Compared to Type I platforms, they also take care of the delivery of the food to the customer. In many markets, these platforms allowed more premium restaurants, often in prime locations, to enter the delivery market without having to build their own delivery operations. Third-party delivery platforms usually charge between 25% and 35% commission.
Type II b — Full value chain models
These platforms are exactly like Type II platforms, but they also operate their own ghost kitchens out of which they launch virtual food brands. The platforms create virtual brands based on the data they collect about their own restaurant partners. The platforms know exactly which type of cuisines and menu items are successful at which price points and in which areas of a city. This gives them a tremendous advantage in creating the menus of new virtual food brands. By preparing the food in ghost kitchens, they also have a cost advantage over restaurants that need to operate out of prime locations with high rents. We don’t see this type of horizontal integration in Germany yet, but it's already common and practiced by the same players that were mentioned above (Deliveroo, Uber Eats) in other markets.
A unit economics deep dive — are delivery platforms really evil?
I am a big fan of looking at numbers and data in order to come up with an opinion based on facts. I am also fully aware that there is a very heated and emotional discussion around whether third-party delivery platforms are evil and are destroying restaurants.
Here you can see an example of a P&L of a restaurant that was presented in a blog post as an example of why third-party delivery companies are evil. The author intends to show you what happens when a restaurant switches from 100% dine-in customers, to 100% third party delivery orders:
As you can see, the author keeps all revenues and costs equal while adding the platform fees to the calculation. You don’t need to be a winner of the Math Olympics to understand that of course, the margins decrease if you add more costs while keeping everything else equal. It's tautological. The author keeps the cost of employees fixed as well while in reality, you don’t need any more waiters if your business is 100% third-party delivery. The author even mentions in his blog post that the numbers are based on his friend’s restaurant who enjoyed record-high sales but he doesn’t adjust the sales in the model.
So, what does this analysis tell us? It tells us that it doesn’t make sense for a restaurant to go from doing 100% in-house revenues to 100% third-party delivery business. But this is not the right approach to determine whether third-party delivery companies are evil or not.
The unit economics of delivery platforms
Now let’s dig a bit deeper into the unit economics of Type I and Type II platforms (Type II B platforms that deliver food of their partners are equivalent to Type II platforms for the sake of this analysis). Please note that the costs and assumptions here are based on a developed country like Germany and need to be adjusted to different markets, e.g., because labor costs relative to food costs may be higher.
You see the unit economics of a Type I platform (grey background) on the left and the Type II platforms (pink background) on the right. Each has two columns — the only difference between the two columns here is the basked size. In order not to bore you with explaining my assumptions (see separate section below), I would like to start with the key take-aways:
- The Type I business is much more profitable than Type II business models for platforms
- Rider costs and utilization are key drivers for platform profitability
- The size of the basket has a massive impact on profitability
- Marketing costs of platforms are high and are heavily affected by the level of competition in the market
As we can see, the direct margins of the food delivery companies are very thin because delivery and marketing costs are very high. These margins do not consider any headquarter and other overhead costs yet. In order for platforms to be able to be profitable at some point, they must scale massively, which explains the M&A activity we see in the food delivery markets around the world, and we are likely to see much more consolidation in the near future.
Especially the Type II model is very challenging to operate profitably. A closer look at the Deliveroo annual report of 2020 shows how difficult this is. In a market that was extremely favorable for them due to Covid, they still lost $ 309 million. Deliveroo is using gig-workers in most of their markets which makes their deliveries comparably cheap as they often don’t pay social security etc. With the government is increasingly pushing back on the gig-economy “employment” model, the stock price of Deliveroo has been under pressure.
In summary, while restaurants perceive the commissions of around 30% for Type II model to be very high, it’s not impossible, but very hard to operate these platform models profitably.
Details & Assumptions — skip this section if you are not interested in the nitty-gritty:
- General assumptions:
All of these numbers are based on what you would expect in Germany and maybe significantly by market depending on the maturity as well as the salary and general cost level in the respective markets.
The assumption is that the delivery platform in this example is the market leader and the market is in a “steady-state”, meaning the competitive intensity is rather low. In a highly competitive market, the marketing costs per order would be much higher for platforms and potentially also the rider cost would be higher due to higher salaries and lower utilization per rider.
- Assumptions per line:
Basket size: The left column has a basket size of 20 EUR, while the right column has a basket size of 25 EUR.
Commission: We assume 13% for the Type I model and 30% for the Type II model because that’s what the largest player in Germany (Lieferando) charges.
Delivery fee: The average delivery fee that is charged is 2 EUR in Germany. In some cities/ regions, this goes up to 2.90 €.
Rider costs:
12 EUR salary per hour (slightly above min. wage in Germany) + social security + insurance + depreciation (e.g., for bikes) + material, etc. = 14 EUR per hour (assuming the rider is properly employed)
2.5 orders per hour on average. Of course, a rider can do more deliveries during peak hours, but given varying utilization throughout the day, even with operational excellence, it would be hard to get above 3 orders/ hour on average, so each hour of the day
Voucher/ Marketing: The marketing and voucher costs of a platform is highly dependent on the competitive situation of the market as well as the stage of the platform. Especially during the high growth phase of the platform, the marketing/ voucher costs of the platform may be significantly higher than the 1.5 EUR that I assume for this calculation. It may also change over time with competitors entering or leaving the market. Moreover, Type II platform usually have higher retention of the user thanks to the offline presence by having riders branding on the streets
Other direct costs: This includes costs like refunds, customer service/ call center, partner handing, etc.
Note: payment costs are assumed to be zero. The assumption here is that the market leader is able to build their own payment processing in order to avoid payment fees for payment gateways like Stripe or Adyen. In reality, many platforms incur direct payment fees which lower the contribution margin.
Unit economics of restaurants: own delivery vs. working with third-party delivery platforms
Now that we understood the unit economics of the delivery platforms, let’s look closer into the unit economics of restaurants (detailed assumptions — see the section below). I want to compare the two delivery options that restaurants have: building up their own delivery service vs. working with third-party delivery platforms.
The most interesting learnings here are:
- The margins of “own delivery” versus “third party delivery” are really similar. The main reason is that deliveries are simply expensive, independent of whether the restaurant does the delivery on its own or via a third party.
- Margins are heavily impacted by the average basket size.
The second point is interesting. The smaller the average basket is, the more profitable third-party delivery is for the restaurant. The larger the average basket of a restaurant, the more it makes sense for a restaurant to deliver on its own, because delivery fees of third parties are usually not capped and the 30% fee is charged independently of how big the basket is. That being said, there are also examples of platforms that cap their fees (e.g., Slice).
Please note that I did not include any marketing costs of the restaurants here to acquire and to keep new customers for the restaurant. If we consider these marketing costs (more on marketing below), then “own delivery” margins will even be smaller. In other words, own delivery makes sense if restaurants are able to generate repeat business from their customers on their own platform, so that the marketing costs for these future orders can be considered to be zero.
Just like it was the case for the delivery platforms, it is absolutely essential for the restaurant to properly utilize their riders also in off-peak hours. Yield Management and platforms like DiscoEat can help to avoid huge demand swings that make capacity planning very difficult.
Good job — you managed to get through the number crunching! I know it’s much more fun to just bash the delivery platforms, but I think it’s important to really understand the economics of food delivery in order to form educated decisions. Independent of whether food is delivered by the restaurant or by third-party delivery companies, food delivery is a business with high costs and, therefore, with very low margins.
However, it’s possible to be profitable with operational excellence and by finding ways to increase the basket size.
Details & Assumptions — skip this section if you are not interested in the nitty-gritty:
- General assumptions: Same as above.
- Assumptions per line:
Basket size: Same as above.
Direct food costs: Assuming a well-run and managed restaurant, the direct food cost (= Nettowareneinsatz in German) should be around 25% for a restaurant. Side note: unfortunately, many restaurants don’t properly calculate this on an item basis.
Rider costs: I am assuming the same rider costs here as for the delivery platforms. There are arguments for and against why a restaurant's delivery operation should have a higher/ lower cost per order which is why I kept the value the same.
Commission: We assume 30% for the Type II model because that’s what the largest player in Germany (Lieferando) charges.
Payment fee: I am assuming a 2% payment fee for the restaurant for “own delivery”. I am assuming 0.59 EUR payment fee for Type II delivery, because this is what the largest player in Germany (Lieferando) charges.
Packaging: The assumption here that packaging (all-in) is around 25 Cents per person, and I assume on average two people per order, so 2 dishes per order.
Other direct costs: This includes costs like refunds, customer calls (for restaurant own delivery), etc.
Marketing and user acquisition
I briefly touched on marketing and user acquisition above, but I believe that we need to elaborate on this topic a little bit more. Both the platforms as well as the restaurants need to acquire users or customers.
In an ideal world, a restaurant has a stable base of regular customers who come back to the restaurant frequently whether it’s for reservations or for delivery/ pick-up orders. The more well-known a restaurant brand is and the more regular customers a restaurant has, the higher the percentage of organic customer growth due to a strong online following and free word-of-mouth marketing. When Covid hit, the restaurants that were the most successful were restaurants that had a strong following and potentially even a CRM system or an email list that allowed them to communicate with their customers directly and to inform them about their new delivery service. Those were also the restaurants that were the least dependent on third-party delivery platforms. Great branding and customer relationship management allowed them to stay in touch with their customers throughout the pandemic.
Like the restaurants, delivery platforms are in competition with each other, and they must acquire new users in order to grow. Because of their heavy marketing spending, they usually need users to come back to the platform several times before a customer breaks even. In other words, the customer acquisition cost of one customer is often much higher than the revenue that the delivery platform generates with their first order. While many users churn after the first order, the platforms need to create a habit for their users to keep using the app. The platforms must create a habit for their users because it makes them very unlikely to switch to other apps or to order directly from a restaurant due to the convenience factor explained above.
Lastly, it’s always cheaper and more beneficial for a platform to acquire new users than it is for restaurants. On one hand, platforms are better and more efficient at marketing. On the other hand, the acquired platform user has the chance to order from many different restaurants, while a restaurant can only offer its own service to the acquired user.
So why does everyone hate third-party delivery platforms?
Looking at the unit economics above it becomes clear that the fees of third-party delivery platforms, while being high in absolute terms, often do not cover the full costs of doing these deliveries. We saw that food delivery usually has rather low margins and even restaurants that operate their delivery service successfully struggle with profitability. In a recent podcast Dara Khosrowshahi, the CEO of Uber, was confronted by the podcast host with the words “They (the restaurants) hate you”. So, there must be other reasons why restaurants hate third-party delivery companies, and, unfortunately, and it’s easy to find many mistakes that these companies have made. Here are just a few examples:
- Shadow websites that seem to be restaurants’ actual websites (e.g., Lieferando)
- Limited to no control over refund process (e.g., Uber Eats)
- The exploitation of riders of the gig economy (e.g. Deliveroo)
- Avoiding workers councils & undermining worker rights (e.g., Lieferando)
- Questionable restaurant support programs (e.g. Grubhub)
- Listing restaurants without their permission (e.g., GrubHub, Seamless, Yelp)
Now it’s clear that these companies have made mistakes in the past that have enraged restaurant owners and they now have to work hard to regain the trust of their partners. Given that many of these companies are rather young and grow extremely fast, it’s also no surprise that they will make mistakes and I am sure that they do regret these mistakes and will work hard in the future to regain the trust of their partners and users alike because no platform can strive if it’s hated by one side or even both sides of the marketplace, no matter how big the venture funding is.
So, are third-party delivery platforms really evil?
As always, the answer is “it depends”. I believe it’s important to find answers based on data and to avoid heated emotional discussion around the topic. Quotes like “In order for DoorDash and UberEats to enjoy attractive profits for themselves, they take them from restaurants.” may be very popular, but they are simply false (both DoorDash and UberEats are still extremely unprofitable). We should have an open exchange based on facts. The reality is — food delivery is expensive. Commission caps that were implemented in some cities do not solve the problem of high delivery costs. In my view, the biggest problem is that customers are not willing to pay the real cost of food deliveries. Often users expect free deliveries as many marketplace platforms used it as a USP in the past to reduce the risk for users to try buying online (see Amazon as an example).
So, in a best-case scenario, restaurants have built a great brand and generate their revenues mainly from their own loyal customer base. A restaurant in a central location with expensive rents cannot survive if they generate their revenues mostly from third-party delivery platforms. The food delivery business is a business with very low margins, no matter whether the restaurant does the deliveries or the third-party platforms. There are cases where it makes more sense to use a third-party delivery platform (e.g., with low baskets) and others where it may not make sense to not collaborate with them. There are certainly issues with third-party delivery platforms that were mentioned in this article as well. An ideological war however makes no sense. The reality is third-party delivery companies exist and many people love to use their service and stay mostly loyal to these platforms. Restaurants can and should add the platforms to tap into the customer base of the platforms and see them as additional revenue channels. If the platforms learn from their mistakes and become more customer and partner-friendly, then I believe we can live in a reality where both restaurants and platforms can successfully co-exist and strive.
Note: This article was written based on a keynote speech that I gave together with Max Kochen, the owner of the German restaurant chain Beets & Roots at Germany’s largest Gastronomy trade fair “IDX by Internorga” in March 2021.
Sources and further reading:
https://www.notboring.co/p/the-beginning-of-the-end
https://thespoon.tech/restaurants-hate-third-party-delivery-services-actually/
https://www.afr.com/technology/why-restaurants-hate-uber-eats-20190716-p527lt
https://edition.cnn.com/2020/11/27/business/restaurant-delivery-alternatives/index.html
