PM home task walkthrough: Uber Eats
An overview of a real-life product management take-home assignment
I love take-home tasks. Unless they take more than 10 hours and I get no feedback. But more often than not, they are great fun and help the hiring team understand how one thinks. Diving into a totally new industry, forced to face ambiguity and make assumptions - a very lifelike representation of a product manager's day-to-day life.
This one here is a home task I created last year. I’ve made some adjustments and added a few comments here and there to help you follow my thoughts better. I hope that by the end of it, you’ve found a few nuggets for yourself or you’ve gone in a totally different direction while thinking along. Latter is totally fine, even expected. I do not consider my solution the right, the only, or even the great way. There are hardly any black-or-white answers for these types of tasks. You’ll enjoy the read more if you follow along. Looking forward to your thoughts in the comment section!
The task
(Obviously, I’ve lost the original, so here’s the gist of it)
You are a product manager at Uber Eats and the VP of product wants you to validate a new business idea that the CEO pitched him. He slacked you this:
“Due to fluctuating demand, there are certain timeslots where restaurant kitchens are getting way fewer orders than they could potentially serve. That’s a huge amount of wasted resources that we could turn into revenue. We could help restaurants maximize their kitchen and workforce efficiency by letting them run Uber Eats branded virtual restaurants on demand. Basically, when the restaurant has extra capacity, it can switch to its Uber Kitchen virtual venue on the Uber Eats app and fulfill orders coming in from the additional venue. Because they utilize our brand, we should get paid for this, so it’s basically a franchise. Figure out if and how we should proceed with this idea.”
Uber Eats virtual restaurant franchise
Overview & background
This analysis evaluates the idea of offering Uber’s very own ghost kitchen as a franchise for existing Uber restaurant partners to maximize their kitchen efficiency. This business opportunity will be called Uber’s Kitchen throughout the analysis and it consists of the following parts:
Overview & background
Involved parties
Business model
Forecasts & targets
Risks and unknowns
Competition & market trends
Conclusion
Ghost kitchens - the concept
Also known as dark kitchens, cloud kitchens or delivery-only kitchens.
A ghost kitchen is a fully-equipped commercial kitchen specializing in producing food exclusively for delivery, without dining areas or customer seating. Think of it as a kitchen space for hire where food entrepreneurs can rent or buy space and equipment for food preparation and delivery.
Grand View Research estimates the value of the global online food delivery market to reach nearly USD 388.74 billion by 2028, with an impressive 10.8% compound annual growth rate between 2022 and 2028.
Virtual restaurants
This term is often used interchangeably with ghost kitchens, but (at least for this context) is better described as a type of ghost kitchen, where an existing restaurant expands its menu for delivery purposes. The important difference is that the restaurant continues its regular operations on the front lines and increases the number of orders through a delivery-only virtual kitchen to better utilize the already existing resources.
Hypothesis
By introducing Uber’s kitchen, the restaurants will be able to increase their kitchen efficiency by activating an additional (virtual) kitchen on their premises on demand. This will also help restaurants to protect themselves against fluctuating demand and earn more revenue with minimal extra investments.
Uber will also increase its revenue through a higher number of orders and franchise fees. In addition, the restaurant partners will likely retain longer with Uber thanks to deeply rooted business operations.
Target market of the analysis (attributes I considered)
Country (population (in serviceable area))
Potential user base %
Existing restaurant partners (of which could potentially use Uber’s Kitchen)
Involved parties
Restaurants
This model would potentially be a good fit for restaurants that are not fully utilizing their kitchen space and staff, are or want to be delivery-focused, see fluctuating demand and don’t have in-house marketing competence.
Requirements:
Internal staff training
Managing Uber’s Kitchen supplies and ingredients
Operating a virtual kitchen through the Uber Eats app
Benefits:
Additional revenue stream
Increased kitchen efficiency
Customers / eaters
Requirements:
More selection (choice overload risk)
(Optional) Uber One (premium subscription) for better prices and/or promotions
Benefits:
More selection (for locations with fewer restaurants)
Trusted quality food
Better prices (either directly or indirectly through increased restaurant efficiency)
Courier partners
Benefits:
More orders with the same amount of locations
Lower chance of running out of orders
Uber
Requirements:
Analyzing data for consumer demands and areas with a lack of supply
Operational resources for creating and maintaining menus
Development resources for building and maintaining tech
Marketing resources for franchise takers
Conducting regular quality checkups for franchised food
Logistics resources for supplying restaurants with ingredients
Benefits:
Additional revenue through franchise fees and increased orders
Increased brand affinity by eaters
Increased restaurant partner retention & satisfaction
Flowchart of the operations
Business model
NB! This model is put together by today’s best knowledge and contains many estimates
Restaurant
Starting / setup fee: X€
Commitment from the restaurant side
On-premise setups and reviews
Putting together a tailored menu + other onboarding tasks
Marketing costs
Average menu item price: 10€
Royalty / franchise fee: 10% of gross sales
Agency / Uber Eats platform fee: 20% - 35%
COGS: 30% - 40%
Other expenses: 5%-10%
On average, restaurants should gain a healthy 20% margin
Average orders per day: X
Average order size: 10€
Average extra revenue per day: Y€
Profit per day: Z€
Profit per month: W€
Breakeven point: 1-2 months (based on my rough calculations)
Uber Eats
Costs: (starting cost + variable cost (per market, restaurant and in total)
Feature launch investment
Development + Further research + Design + Operations
New market launch investment
Marketing + Operations
Restaurant launch investment (covered with setup fee)
Onboarding and support
Variable costs
Development and maintenance + Marketing + Data analysis + Logistics +
Per restaurant
Logistics + Quality assurance + Support
Revenue:
Royalty / franchise fee: 10% of gross sales
Average restaurants using Uber’s Kitchen
Average orders per day per restaurant - calculations using:
Average order size
Average monthly revenue per restaurant & market
Average monthly profit per restaurant & market
Forecasts & targets
Based on the rough calculations in the previous chapter, it seems to be possible to make this model profitable. Due to many assumptions, the confidence is not high enough to immediately launch this opportunity. To raise confidence in this bet, further research would be needed, mostly regarding demand verification and feasible logistics & pricing models. The decision to go or not go for this opportunity also depends on the business goals and comparisons with other opportunities.
In any case, on a single market, Uber’s Kitchen would hardly break even and not justify itself. On approx. five similar size markets, it would take around half a year to break even and then start making a profit of around X€ per market. (these calculations were done based on the assumed numbers I didn’t include)
Risks and unknowns
Risks
Drifting away from the core business
The further we go from the core business, the bigger the risk of maintenance costs eating away all the upside. Multiple new operations would need to be started here, e.g. creating & updating menus and planning & delivering ingredients to the restaurants.
To minimize this risk, the target market size has to be validated and big enough to justify itself. New operations should be covered with as many existing ones as possible, e.g. combine Uber Eats Market logistics with Uber’s Kitchen logistics as much as possible.
Lack of advantages
To compete with other restaurants in the food delivery market, the food should stand out by being cheaper, better or different. Too cheap would manipulate the market and work against us in the long term, better food would be hard to pull off without additional expertise, but different could involve a marketing opportunity that needs further research.
Assumptions and best bets
Pricing
Simply going with lower price points than the market would likely result in unequal opportunities from the restaurants' side. To avoid competing with restaurants, there has to be an incentive for them too. Lower than market price with Uber One -> more orders for restaurants in general (because they are committed to Uber Eats over competitors -> losses from lower than market price will be compensated by additional orders by Uber One customers -> bigger incentive for restaurants to join Uber’s Kitchen franchise.
Logistics
Getting restaurants the ingredients they need when they need them is potentially the biggest make-it-or-break-it factor in this formula. The idea of a lightweight menu that serves as an extra line for the restaurant means that the volumes would be rather low. Low volumes drive logistics costs high. Two possible ideas on how to reduce this risk:
Uber Eats suppliers deliver supplies to Uber Eats Market locations, where Uber Delivery Partners deliver them to franchisees. Due to additional development costs and low reliability, there’s a high chance it won’t solve the problem.
We let restaurants take care of getting the necessary supplies. Having them be in control of the whole supply chain reduces complexity, but also increases the risk of food quality being too unpredictable. This risk can be reduced by varying menus based on the restaurant's specialty and researching top suppliers' offerings. Another way to reduce risk here would be to just start with a few “Uber collab” dishes in the restaurant’s own menu.
Conflicts of interest
Uber - Restaurant - Cuisine
If Uber would franchise foods from the same cuisine as the serving restaurant, chefs would be more skilled in the dishes, but these dishes could steal revenue from the restaurant’s own menu.
If Uber franchises foods from different cuisines in the serving restaurants, chefs would be less skilled in the dishes, but the “serving spectrum” and number of orders would increase more.
Restaurant types: traditional vs ghost kitchen
Not fully utilized kitchen space, which is one of the problems being solved, is typically an issue for traditional restaurants focused on in-house dining. Yet, traditional restaurants are less likely to run a virtual restaurant from their premise.
Competition & market trends
Direct competition
There are no examples of food delivery services offering lightweight virtual restaurant franchises.
Wolt has a few examples of collaborating with restaurants on a single-dish basis, but this hasn’t seen much traction or scaling. NB! Their goals with these kinds of collaborations are unsure.
Wolt also has examples of promoting special offers for subscribed users. This approach has seen some decent traction based on its subjective growth so far. Subscription service itself has proven highly successful in 22 markets, with 70 million+ saved for users. (source)
Indirect competition
On-demand host kitchens offered by Deliveroo (source)
Deliveroo has created ghost kitchen “hubs” that they rent out to businesses who want to run virtual kitchens without any hassle. The package includes rent, equipment, waste management, marketing and a site manager. The fact that they started already 8 years ago with this concept and have scaled it to five different countries already shows that there’s at least some sort of traction. They seem to mostly target areas that naturally don’t have many restaurants to increase the coverage.
Overpopulation of ghost kitchens (source)
It should be kept in mind that due to its scalability, the ghost kitchen concept can be abused in a way that hurts the competition in a non-healthy way. The same could happen with Uber’s Kitchen if it gets top results by default and halts competition with unsustainable offers for restaurants.
Conclusion
While the concept of offering Uber-branded virtual restaurants as a franchise can be innovative and helpful for struggling restaurants, there are many unknowns and assumptions combined with no existing examples of such business models. With current knowledge, there is not enough confidence that the high risk of the idea would justify itself. The recommendation for the next steps is to validate the riskiest hypothesis with the lowest risk proof of concepts.
Proof of concept experiment: “Promo dishes”
It is essentially a 20/80 of the original idea: a low-risk, low-investment experiment that will help validate the riskiest assumption to decide if this idea should be greenlighted or not. The idea is to partner with restaurants to sell Uber-branded dishes, like Uber Pizza or Uber Burger, through their existing venues. Where and how to do this needs further research, but it should be tested on multiple markets and restaurant types and cuisines for sufficient confidence. These are the risks that will be validated with this experiment:
The risk of insufficient demand
This could come from restaurants not being interested or not having the capacity for / wanting “a kitchen inside a kitchen”. It is also not certain that there will be demand from the eaters’ side. Both of these biggest assumptions can be tested with the “promo dishes” experiment.
The risk of logistics complexity
In this experiment, different logistic approaches can be tested. The most far-fetched one is the one where courier partners would move ingredients between Uber Market and restaurants. The feasibility and cost-efficiency have many unknowns that can only be solved with real-life testing.
The risk of pricing
The difference between market or below-market pricing effect and/or combining offers with Uber One subscription has also many unknowns. A lot more confidence would be gained through “promo dishes” experiment.
Success
We will be ready to move on to the franchise stage testing after seeing successful results for these three experiments on different markets:
Shipping our own ingredients to restaurant partners so they can cook and sell Uber-branded food has helped restaurants increase demand over an extended period of time (1 month+). This also validates the eater demand.
Finding a profitable way to ship our ingredients to these restaurants is a critical experiment, too. The model shouldn’t add more than 10% to the ingredients price on a large-scale operation.
Selling our branded food at market rate is part of the demand verification. Combining it with Uber Plus offers will help us understand if that could be a way to increase demand without hurting our partners. For that, the additional revenue from Uber Plus should be at least double the amount we spend on the promotions.
The main problem behind the scene seems to be the peak times of orders, which are creating overload and inefficience for restaurants (where this idea comes in as a solution).
How is it possible that creating the ghost restaurants will change the peaks and relieve the fluctuating overload, caused by user demand?
Why should the users change their behaviour and start to order off-peak time ? Better price? Logically, this will create extra demand for peak times instead avoiding them.
Also, why don't the restaurants use this idea already and do not relieve the pain of temporary overload by offering lower price at off-peak times?