The problem with the delivery app business.

 

The problem with delivery apps in India

When first studying economics and later business, I learnt that building a profitable business depends on: Supply and demand (if demand goes up relative to supply, price increases) & Solving a problem better than competition.

The example of Uber and Ola in India illustrate what I mean by supply and demand. Ride hailing apps worked well in markets like the US where they got car owners (e.g. retired people or college students)
to work part time as Uber drivers. This added to the number of cabs available. When supply went up, relative to demand, prices fell.
The reduced prices attracted new customers, who wanted a better alternative to public transport or their own vehicle, but at a lower cost than conventional cabs. The lower price was acceptable to the supplier (Uber driver) as it was extra income for him, thru utilizing his idle asset (car) and time.

In India, the average car owner would not want to drive a stranger for money and lower himself to the status of a taxi driver instead of a upper middle class car owner.  Thus, an Uber/Ola driver in India is simply a new kind of taxi driver, paying for a new vehicle and doing his job full time. His income expectations are similar to a regular taxi/auto driver.  The only way to stimulate additional demand was for Uber and Ola to discount riders (while incentivizing drivers).  
The discounting was unsustainable beyond a point. As fares have increased to the same level as a regular cab or auto, many existing auto or taxi drivers have used the app as a means of arbitraging fares.  There are genuine problems that the app solves (like using it when you are unfamiliar with the destination) which has led to a modest increase in demand (enough to sustain the additional supply of drivers).

Margins of the aggregator will drop once more apps (e.g. govt sponsored apps with no profit motive) are offered e.g. Namma Yatri in Bangalore. those on apps will game the system (cancel the ride and pay the driver outside the app). Those not on the app will improve services (using GPS for e.g.), all of which leads to normal profits and not projected high margins, justifying high valuations.

This is the context in which I’d like to look at food & grocery delivery aggregators.

Home delivery at short notice has been around forever. If you called your grocery store, a delivery boy would come to your home and no extra payment was expected. It worked because the retail business, which remains mostly unorganized, is disguised unemployment. There is no extra manpower cost and transport cost (in densely populated Indian cities, most restaurants and retail outlets are less than 2km away) is negligible.

Apart from offering home delivery, the apps offer the customer discovery. In theory a customer who was not aware of a restaurant listed on the app, can now order from it – particularly when the restaurant is further away than the distance the customer is prepared to travel to, or the restaurant is comfortable with, to deliver themselves. The problem is that when everyone is on the platform, the restaurant is as likely to lose customers to newly discovered restaurants, as it is to get new customers. The only differentiator is price. When the app can no longer afford to finance discounts, partnering restaurants must do so, leading to a race to the bottom on margins.

Although home delivery aggregators offer, in theory, an additional way to consumer the product, which should lead to more sales per restaurant, the restaurant business in India has only grown as the same rate as India’s GDP (9-11% p.a inclusive of inflation). The rate of growth of the restaurant business (adjusted for COVID) was the same in the period before and after delivery apps.

If one looks at the same store sales of restaurants in listed companies, Specialty restaurants had a per restaurant sales of 2.36 crore p.a for f.y 2012 and sales of 3.08 cr in f.y 2023. The increase of 2.5% p.a
is less than inflation.  India’s largest chain Devyani foods has a decline in per store sales between 2020 and 2023. At the same time, off premises sales have increased for all restaurants. This trend has sustained after covid. Thus, while sales per store (adjusted for inflation) remain constant, all that has happened with the advent of delivery apps is that the sales have been split between dine-in and home delivery. The latter involves a commission to the aggregator, which erodes margins. Zomato’s income is around 24% of order value (net of GST).

Although some of that 24% is a delivery cost charged to the customer, the restaurants, which operate on a EBIDTA of under 20% cannot afford to give away even half of this in commission to the likes of Swiggy and Zomato without seeing any incremental sales volumes. 

Since the price the restaurant charges is a function of their ambience and dining experience, as well as good and home delivery does not provide the out of home dining experience, there is more resistance to pay the restaurants listed price. While a restaurant would ideally like to increase prices to cover the commission to aggregators, the customer wants to pay less. This has led to the rise of cloud kitchens – which offer home delivery only food at lower prices than a full-service restaurant and apps like ONDC which charge negligible commissions, enabling the restaurant to pass on some of the resulting savings to the customer. 

Zomato reports a monthly active user base of 16.6 million, which has remained fairly constant in the last 2 years. They list 215000 restaurants active each month, with 316000 delivery executives. This number of users and restaurants is pretty much the market size in urban India. This is after Zomato withdrew from 225 towns which contributed negligible sales. This base overlaps with Swiggy.  

 The average order value for Zomato is listed at Rs 408.
If there are 16.6 million transacting customers per month and each orders just once that month, either the average order value will be Rs 330 (incl tax), or the number of customers has been overstated. 

About 10% of customers are members of Zomato gold (1.8 million at the end of Mar 23). Assuming an
average monthly purchase of Rs 1000, the average order size of the remaining 90% would be Rs 342
(incl tax) per month.    

Although a lot of startups, particularly Unicorns justify their valuations based on the 1 billion+ Indian consumers, the reality, as seen across product categories, is that a max of 20 million Indians can buy consumer products online and 90% of them spend (in this category) under Rs 250 a month. This is not an amount that translates to a profitable customer.
There are an estimated 24 million adults in Urban India who are potential customers for regular online purchases. Zomato/Swiggy already access most of them.

The 315000 delivery personnel of Zomato of an average of under 2 deliveries a day (Rs 660/day).
Assuming the same workforce also delivers groceries (the Zomato hyperlocal business), it is still
under 4 deliveries a day. Even if the delivery person is listed on 2 platforms and does 8 deliveries a day, his net income is barely half the minimum wage.

For an average restaurant partner, Rs 25500 (incl. tax) per month is the value of business from Zomato.
At an average order value of Rs 408, this is 2 orders per day, which the restaurant can execute themselves with their existing staff.  Swiggy has lower business per restaurant.

After 15 years of operating in India and having accessed most of the total addressable market, Zomato still makes a loss at EBIDTA level. It is the same with other companies in this category. I have quoted Zomato as it is a listed company with publicly available financials. Swiggy’s losses in f.y 23 were more than half its turnover.

Going forward, the problem for Zomato/Swiggy is that they have already reached most of the potential market – the cost of acquisition of incremental customers may be more than the value from them.
They cannot increase commissions from restaurants any further – with ONDC there will be pressure to
reduce them. ONDC prices show the customer that most restaurants are increasing the price for online deliveries through Zomato & Swiggy. The customer will not pay more and the main element of cost (delivery boy fee) cannot be reduced as they already earn very little in absolute terms.  That is a conundrum that the aggregators will find difficult to resolve.

One opportunity is for aggregators to use the data they have, to integrate backwards into food supply, using dark kitchens, with a better product offering than restaurants that lack data on customer preferences. However, both Swiggy and Zomato have got out of this business. The companies that are offering a dark kitchen, delivery only model (Rebel foods, @Kitchens etc.) are unprofitable.

The same kind of problems exist in the grocery apps (hyperlocal instant delivery).
I had argued years ago in my LinkedIn article
https://www.linkedin.com/pulse/coming-meltdown-indian-e-grocer-business-model-rahul-deans/?trackingId=LV8Rsj2%2BTfaoEAE%2B49j0RA%3D%3D that the online grocery business was flawed. Since then, the online grocery companies have all merged with offline ones (Amazon acquiring More, Tata acquiring Big Bazaar and Reliance with Jio Mart) to be integrated retailers, since online only was not a viable strategy. None of these online only companies was ever profitable).

 Both Zomato and Swiggy got into hyperlocal, in my view, not to solve a customer problem, but to better utilize their delivery staff. Armed with other people’s money, they may also have felt they could do a better job than those that tried earlier. Many of the users of hyperlocal apps are driven by offers and discounts. e.g. if I get a free coke with my first order on Blinkit (Zomato’s grocery delivery app) and I am going to buy those items anyway, I will use the app – and use it again only when I get a similar offer.   

The quick commerce business of Zomato had a loss of 562 cr on orders worth Rs 806 cr that were delivered, in f.y 23. The average order value of Rs 522, is just above the threshold required to avail free gifts. At an average gross margin of at least 21% (which is what a Supermarket gets), Blinkit should have a gross profit per order of over Rs 100 (excluding GST on the order value). After considering delivery costs and operating costs of a dark store (doing 2 crore / month), it should have a contribution per order of Rs 30. Instead, it has a loss of Rs 14 per order. Instamart’s losses are expected to be higher, while the third player in this space, Zepto had losses in excess of turnover in f.y 22 (fy. 23 results not yet announced).  

My other posts on startups in this blog 

https://rpdeans.blogspot.com/2023/11/joining-startup-checklist.html

https://rpdeans.blogspot.com/2023/08/the-coming-unicorn-meltdown.html

https://rpdeans.blogspot.com/2023/07/why-coffee-chains-are-still-not.html