What’s wrong with food tech?15 min read . Updated: 17 Mar 2016, 01:20 PM IST
A geeky chat on the business of food tech
A geeky chat on the business of food tech
Mumbai: This conversation is off the record. Yes. Off the record.
It means the person talking here doesn’t want to be named because he doesn’t want to fall foul of the company he works for—a start-up, where he is part of the top management—where a good amount of investor dollars are riding on him and his team. There’s more to his request for anonymity. Food tech has had it pretty rough the last few months—what with several start-ups shutting shop, others scaling down and a few others rethinking whether there’s money to be made in doing whatever that they do. So, with everything that’s happened, nobody wants to indulge in schadenfreude on what’s going on.
But then, one must take stock. And put things in perspective. This is that story. With a man, we will call Mr. A, who has 5+ years of experience in the food tech business. Edited transcript follows. [Some sentences have been translated from Hindi to English]
Mint: The way it used to work is, you call a restaurant, place an order and done. Your food arrives, maybe 10 minutes late but it does. Where’s the business case for a food tech company here?
A: So, let me be honest. We are not solving any problem here. Ordering through a phone has never been a problem. You have five restaurant flyers at home, you pick one and order. At times, the restaurant might even remember you and know your address and deliver the order faster. The problem has always been at the restaurant owner’s end. If his delivery business is picking up, he thinks, let’s do more delivery. So what does he do? Marketing—that’s problem no. 1 because the only way of marketing for him is pamphlets. That’s it and you know, that doesn’t work very well. Now, problem no. 2: Scaling delivery for a restaurant is tough. He will have two phones, at best two guys taking orders. Phones will be busy. Imagine writing down orders and addresses and sending one copy to the kitchen and then the order gets prepared and then dispatched. Now, imagine this at peak time, 8-10pm. Now, problem no. 3: restaurants don’t think of this as a service; so packaging and promptness in delivery is a problem. Service varies between a Tuesday and a Friday. All clear till here?
A: Now, market research shows that a good restaurant cannot pass more than 70-80 delivery orders in a day. So, a person manning the telephone can’t pass more than 13 orders an hour. Max: 15. This is where food tech companies come in.
Mint: So, your business model exists because restaurants don’t do a good job of deliveries?
Mint: Anything at the consumer end?
A: There is some convenience, like paying online. Then, a wide range of restaurants to choose from, so the choice of cuisine is higher, so you can discover more restaurants. Plus, there is better explanation of the dishes in the menu. But we haven’t solved the payments problem. Credit cards and wallets have. On the discovery part, all we do is list the restaurants. So, we make a cluster of 3km driving distance and list some restaurants. That’s not very difficult. Zomato does that so well. They list all the restaurants, with reviews and everything.
See, I’ll be honest. IRCTC, MakeMyTrip and BookMyShow—they solved a real problem, of buying physical tickets standing behind a counter. If we don’t exist, or don’t exist in a very efficient sense, where we have a killer, seamless product both from a consumer and restaurant point of view, a person will still be able to order food. Ordering food is not a problem.
Mint: Okay. So, what really is the food tech business then?
A: Simple. It is a commission business. You charge a commission from the restaurant for giving them business. Commission can vary from as little as 10% to 15%, depending on the deals you strike. With bigger Quick Service Restaurants (QSR) & chains, the commission is lower. With individual restaurants, it is higher.
Mint: So, how much does that translate to?
A: Let’s say that your average order value is ₹ 300. You make anywhere between ₹ 30-45. That is the primary revenue stream.
Mint: That’s not a lot of money, no?
A: It is a good margin. It gets worse as the order value goes down. And that’s where the business model challenge comes up.
Mint: Okay. Let’s start with the customers first. How do you get them to come on the website or install the app?
A: See, there is no secret here. Marketing. All sorts of marketing. Digital marketing like Google’s pay per click, search engine marketing, putting up a brand logo inside a restaurant which gives direct brand recall, TV campaigns, radio campaigns and discounts are all ways to get a prospective customer on the platform. Remember, this is all in the hope that people who land on the page will actually convert into orders. That doesn’t happen a lot. So, if 100 people land on the website of an aggregator, thumb rule is 30 bounce. Out of these 70 guys on the page, they will scroll to look for their favourite restaurant...
Mint: Hold on. Have these people already decided what food they are going to eat? Indian, Chinese...
A: Thumb rule, yes. Now, if they are scrolling and don’t find their favourite restaurant, they will look for a substitute. If they don’t like that, they’ll bounce. If there is intent to buy, the person selects the restaurant, selects the food and starts making a cart. If the price is very high, then chances are the customer will bounce. Industry understanding is that out of the 70, only 30-40% convert. So let’s say, only 30 people end up ordering.
Mint: What about apps?
A: Unspoken rule in the industry: Of the total number of people who install your app, only 25% are active. Rest are inactive. Some wake up to order only when there is a big offer but they are not sticky at all. So, on the consumer side, all of this can be clubbed under cost of customer acquisition. If you are running a marketing campaign and giving discounts, then the cost of acquisition can be very, very steep. Almost to the tune of ₹ 300. At worse, when you also have a huge call centre running to pass orders, the ones you receive online to the restaurant, cost of customer acquisition can be as high as ₹ 400.
Mint: Alright. How do you get restaurants on board? What’s the pitch there?
A: So, my pitch to a restaurant will be—the restaurant has this attitude that I don’t need you, I do delivery and I am doing pretty well—if the business of this outlet would have been amazing, you would not have started a delivery business. Restaurants start delivery because they want to optimize the same kitchen. They know that utilization is not good. And your dining business is limited to the number of seats you have. Data suggests that a good restaurant can turn around one table 7 times. 5 times is decent. 3 times means you are going to die soon. Now, how do you promote your delivery business? You get dining business only because you have been around for some time. Then I say, there is a limit to which you can advertise and flyers don’t work. Then I start my sales pitch—that once you are online, people will review you and you will gain credibility, plus you will still be delivering in the same 3km but quite likely you will have more orders. Another sales pitch—since it is online, taking an order is not a hassle. The pitch is that your restaurant will be optimized.
Okay. Simply put, the pitch is, come, list on my platform and grow your delivery business. It is a commission-based model, so why not try it.
Mint: So, how do you do that?
A: This is the difficult part and should be explained with some historical context of how the food tech industry has grown.
Till about three years back, one had to explain to a restaurant that coming online makes sense. Life was not easy. Food tech was rudimentary. Everything used to happen over the phone. Food tech 1.1 was a call centre where you would take orders on the phone and pass it to a restaurant on the phone. There was no value-add. Order passing which was a one-time job, a customer calling a restaurant. That became a two process job. Someone ordered a chicken lababdar, you heard something else and passed it on. A South Indian call centre would have immense difficulty in pronouncing chicken stroganoff—this system was rife with loopholes. Then, in peak hours, call centre numbers would get busy, executives get busy and the passing-on gets delayed. At times, the order passing used to happen after 45 minutes. Now, imagine the customer experience. It was completely manual. Plus, there was no online payment option; so, there were a lot of leakages and money wasn’t coming back. You passed an order and wrote it in a book for reconciliation. At the end of the month, the restaurant would straight away deny the existence of any such order. Or he would say, the customer cancelled the order. Or that he went to deliver and there was nobody at that address.
If you add cost of customer acquisition, staff cost, call centre, companies were actually burning ₹ 300 in processing one order. The cost of passing the order alone through the call centre was ₹ 43. Think of it, cost of customer acquisition, cost of taking the order, cost of passing the order and then cost of collecting money.
Food tech 1.2 was taking order online and passing it through the call centre. That was happening till last year. Again, this doesn’t support unit economics. Food tech 1.3 is complete automation—the order goes through seamlessly from the customer to the restaurant. The restaurant prepares the order and delivers it within 30 to 45 minutes. That is the ultimate goal of an aggregator.
Mint: That’s not happening right now, is it?
A: Let’s just say, start-ups are trying to get there. Everybody is trying in various ways.
Mint: Okay. So, how do you make money?
A: On the product side, the Holy Grail is complete end-to-end automation. No manual intervention. No call centre. On the business side, a lean organization...
Mint: What about discounts?
A: See, in the Indian context, discounts are necessary, to get more customers to your platform. Believe it or not, but if you stop discounts, then there is almost a 40% drop in orders. Discounts work. Period. The trick is how long can one do this so that the customer becomes sticky and used to the product. Of course, you can’t go berserk. Do that and you will soon run out of money.
Mint: Ok, so how do you make money?
A: Commission. That’s the primary revenue stream. This is a difficult question to answer because nobody is making money right now, but various companies are trying various things. So, for example, one particular start-up is now getting restaurants to fund discounts. The pitch there is that discounts work. Now that your kitchen has become efficient and your fixed cost pretty much remains the same, except if you are adding more delivery boys on the payroll, you should offer some discount. If you don’t, your neighbour will. So that’s one way. The other is, paid listings. So, some start-ups are offering the first few results. If a customer is searching for a restaurant in his area, you ask the restaurants to buy the slots. Zomato, Foodpanda, all of them have been doing it. So, imagine if you are selling first few restaurant listings and you have fifty 3km clusters in a city, that’s a decent revenue stream.
One thing that TinyOwl did which was good was incentivizing restaurants for faster delivery. So, if a restaurant was able to deliver food within 45 minutes, they would not charge commission. If delivered in 30 minutes, they would give ₹ 50 as incentive. That was a clear differentiator. That’s a good model which will lead restaurants to focus on delivery. Unfortunately, it didn’t work because there were not enough orders to complement this offer. If it had run this for a year, it could have become a clear differentiator. And then once the volume kicks in, you could pull back on the incentives.
Mint: Ok, so how do you make money?
A: Fuel discounts to start with. Get orders. Do smart marketing and brand investments. Once your volumes are high, ask the restaurants to fund some discount. Then, look for other sources of revenue, like I just mentioned. Remember, all of this can happen, only if your orders are high. On the product side, automation, on-time delivery and a very strict review system. Trending dishes, trending restaurants, that’s next-level stuff.
Mint: How big is a cluster?
A: For a restaurant, his universe is 3km, driving distance. Because you can’t deliver within 30-45 minutes outside 3km, however hard you try and I am talking about peak hours.
Mint: In a distance of 3km, how many restaurants are there?
A: Depends on the market. Delhi is a cluster market. Mumbai, you will find restaurants on the ground floor and then residential apartments above it. Delhi, you will have a society, with a market and all restaurants will be in that market. Pune is spread out like Mumbai. Bangalore, restaurants are on highways. There is a cluster concept in this. With exceptions. Like if a restaurant is 5km away from Hiranandani in Powai, he will still cater to it because it is a huge demand generator. Plus, he will batch orders. Clusters are generally 40-50 restaurants. DLF Phase 4, South City, Safdarjung, Panchsheel, Rajendra Nagar, these are all clusters.
Mint: What is the average order value?
A: Depends. Gurgaon is typically around ₹ 480. Mumbai ₹ 430. Pune is around ₹ 280. Bangalore is ₹ 350.
Mint: Should a food tech start-up do delivery? Own delivery?
A: Ah! This is interesting. So when you go to investors for funding, they will tell you that in the food business, there are 3 things on which a company needs to have control. No. 1: sourcing of raw material. No. 2: cooking great food. No. 3: delivering it on time to the customer. And then they will ask you, if you are an aggregator, you are not doing any of the above so what value-add are you bringing to the table? It is a good question but I think there is value in being an aggregator. And what you bring to the table is technology to optimize the efficiency of a restaurant. And a hassle-free experience for a customer to buy food online or from an app.
I think food delivery is a separate piece altogether. There needs to be someone like a Roadrunnr for that. But if an aggregator is doing delivery, then the only thing is that the average order value should be high. You cannot own delivery and be delivering food of ₹ 150. I would say that the order value has to be in excess of ₹ 500. With the current salary of delivery boys, which is now ₹ 13,000+ per month, the cost of delivery per order is at least ₹ 60, if not more. So there’s cost of petrol. Plus incentives, around ₹ 1,000 per month. The whole delivery business is a peak-hour game. The guy can do max of four orders in peak hour, if orders are completely optimized. Which means, he delivers comes back, delivers and comes back. Then, there are more variables—like in the afternoon, orders come mostly from commercial areas and in the evening from residential areas. I don’t think one can do delivery everywhere and for every order. If it has to be done, it has to be done smartly.
So, Swiggy has this integrated in its app that if a delivery boy is not free, the order will not go through. It is a good thing but then you lose orders and it is a costly affair if you are delivering for every order. Like, I’ve seen people order for as low as ₹ 70. So, this is a tough one. But if you are doing delivery, then you can control the last-mile experience and then your re-order rate should be good. That is the rationale there, that if you are controlling the last-mile experience, then the re-order rate should be as good as 60%. So, you have a lot of repeat customers.
Mint: Will customers pay delivery charges?
A: It has to be smartly positioned. Like how Amazon does it. You put it upfront, next to the restaurant listing and customers will bounce. Show it, once the cart is ready, then perhaps, the person will go through with it, since he has taken so much effort to select and prepare the cart. The price you charge is also important. The understanding is ₹ 20 as a delivery charge works. Strangely, there is a drastic fall at ₹ 30.
Mint: What are the kind of order numbers some start-ups are doing?
A: Last I checked, which is about a month back, Swiggy was doing some 18,000+ orders a day. Zomato was doing some 12,000+ orders and Foodpanda some 23,000-odd. Scale, in the sense, the number of cities each of them are present in, is different.
Mint: So, what really is wrong with food tech?
A: Depends on who you ask. A lot of what has happened is because too many players jumped in. There was too much competition, which led to discounting, signing up restaurants at low commission, sometimes at no commission—everything was going on. I think now, some sanity has returned to the space. See, there is a market for an aggregator. People are eating. People are ordering. It is not like that has stopped. All the ticketing guys made money because they solved a real problem. Here, you have to create an experience that the consumer finds real value in buying online. With the number of people coming online and the size of the country, this is a huge opportunity.