Get best of the week

As we for a week without money, connections and software launched a food delivery service and almost did not screw up

My name is Dima, I am a student of Baumanka and a programmer with entrepreneurial experience. Together with the guys, IT specialists and analysts, we came up with Quicq - a single urban logistics service that any company can connect to and eliminate the need to maintain its delivery service. Like a fax, only for physical goods. For a week we launched MVP in Yaroslavl, spent 15 thousand rubles at the start, completed about 250 orders and delivered food for 175 thousand rubles. Now the service works at a loss, but we know how to fix it. We write our own software, prepare for testing new hypotheses, and work on bugs.

I’ll tell you how we went from idea to launch - we groped for a niche, developed a business model, looked for an approach to customers and built up work with couriers. Perhaps our experience will be useful to those who face a similar task, and to those who are interested in experiments in the field of unit economics.



Regional food tech - how it all began


The segment of local urban delivery in Russia has just begun to take shape, and under quarantine conditions its importance for businesses and people has grown significantly. Now this is a basic necessity, along with electricity and the Internet. Delivery in general is a large, technological, interesting and rapidly growing market. We set a goal - to find imperfections in this market and create a business there.

On February 1, we had on our hands a list of 11 hypotheses that can be used to launch a startup. We began to dive into the field of logistics with the delivery of farm products. We thought that people have a problem to buy products, and farmers to sell. To confirm this hypothesis, we conducted a dozen interviews with farmers and buyers. We found out that there is a problem to sell, but no problem to buy. Half of the cost of a farm product is accounted for by logistics costs: two farmers living in neighboring villages deliver their products on their own machines. It would be more logical to combine efforts, but for some reason they do not.

We met with the founder of the service "Eat Village" and offered to find together a solution that allows you to organize logistics twice as cheap. Having delved into the details of the process, we realized that these guys cannot be saved by our forces, and the farmers' market is too tough for us. Then we thought about delivering food from restaurants. There is no point in competing with food tech giants in Moscow, so we decided to hold CustDev in the regions. We started with Yaroslavl, the city in which I grew up and I know like the back of my hand. For three days I talked with five local restaurateurs, and during each conversation I heard: “Of course, we need this! Take the money and do it. ” We realized that we had a problem.

We met as a team and thought: “Guess that's good, CustDev is a tutorial, but let's think about how we can put together a project and launch MVP in a week.” So we agreed that we will start next Monday. We divided the main blocks of work between team members: customer search, development of a client application, search for couriers, organization of a courier management system. I was actively involved in sales and was responsible for synchronizing the entire process. But I think my main task was to create conditions under which we would have no chance not to move forward. As a result, the next Monday, exactly one week after the meeting, orders were sent from us.

Business Model: How It Works


In small towns, the restaurant has two ways to deliver food to the client: through aggregators - Yandex and Delivery, or through its own delivery service. Its delivery service costs the company about 130 thousand rubles a month. In Moscow, restaurants basically refuse their own decision in favor of aggregators, but in the regions, in particular, in Yaroslavl, this does not work. And that's why. The radius of delivery of the aggregator is 3 kilometers, while it takes 30% of the check. In a regional city where all the restaurants are located in the center, a radius of 3 kilometers cuts off the bulk of customers living in sleeping areas.

As a result, restaurants organize their own delivery services. Your service can carry food anywhere, but this is a huge layer of additional work. For the price, this solution comes out a little cheaper, many do this: the restaurant leaves its offer in the aggregator, indicates the delivery radius - the whole city, collects orders and distributes them independently. Our business model follows from here. We provide the restaurant with a boxed solution - a delivery service on the principle of a classic outsourcing. The client connects to our service and gets access to your personal account on the Quicq.ru website. The restaurant administrator enters all the information about the order and presses the “deliver” button, similar to calling a taxi. Creating an application on the site takes one and a half minutes. After 15 minutes, our courier arrives, picks up the order and drives to anywhere in the city.



A week our service costs the client about 10 thousand rubles, provided that the restaurant has its own couriers. If there are no couriers, but there are many orders for delivery, the restaurant spends 30-40 thousand a week for our services. We earn by receiving money from the restaurant for each delivery. We do not take a percentage of the order, because we do not care what to do, whether it is a bar of gold or a cutlet. It all depends on the mileage. Now we are experimenting with tariffs, the first week we worked on a fix and delivered for 160 rubles to anywhere in the city. The economy in this case will converge if you send couriers evenly - both far and close. At this rate, restaurants covered close distances with their couriers, and we were sent to distant lands. For some customers, we created a price scale that varied at intervals: up to 5 km, 5-10 km, 10+ km, etc.e. Now we have plans to test the model with payment for each kilometer.

We consider Dostavista express delivery service to be our main competitor. They offer similar services, do it a long time and on a large scale. And yet we have different business models. Dostavista works like a taxi with a focus on B2C, connecting free couriers and customers. If I, in self-isolation, want to send a friend a phone charge, I will definitely do it through Dostavista. If I am an online store that needs to send some goods to the other end of the city during the day, then I will also most likely turn to them. And if I need a solution for local high-frequency and high-speed deliveries, then this is about us.

It's too early for us to compete, but it's time to get inspired and adopt the best practices. Inspired by Dostavista, we began to study the experience of creating a platform, marketplace. And Yandex.Lavka inspired us to think about launching product delivery in the regions.



MVP in Yaroslavl: “if we survive here, we will survive anywhere”


Yaroslavl - this is 600 thousand people, low solvency, terrible roads, heavy traffic jams during rush hour. We chose this city to launch MVP according to the following principle: "if we survive here, we will survive anywhere." By the same strategy, I grew up in it. We decided that if we organize the logistics system so that the economy converges, it will be much easier for us to scale to other cities.

We came to city restaurants and said: "Hello, we are Misha and Dima, we want to solve your problem with delivery." In some cases, it worked, and we scored a pool of clients to start, but then we ran into a fundamental restriction on working in this mode, since most companies are still not ready to work with Misha and Dima, they need a legal relationship and an agreement.

I’ll tell you how the work with couriers works. We searched for them using ads on Avito. We gathered a large number of people with a personal car in the database, divided the working day from 11 to 23 hours into three blocks and formed shifts of 4 hours, in each of which or all at once the couriers are recorded at will. By the beginning of their shift, couriers arrive at the collection point in the city center and await the order. The new order is displayed in the courier’s mobile application, he quickly picks it up and takes it, and then returns to the collection point. Until the process is automated, we take all the work on the formation of shifts.



We realized that it is wrong to pay couriers on orders. To save people, we switched to the unprofitable model so far - we pay the hourly rate, compensate for gasoline and pay extra motivation for each completed order. Our courier earns about one and a half thousand rubles a day, we transfer money to the card once a week. It is important to note that regardless of how many orders we carry out per day, 5 or 25, the cost for couriers is approximately the same, we still put them in shifts and pay them the rate.

If the courier screwed up, the restaurant calls our shift supervisor and we solve the problem. To date, we have completed more than 200 orders and have never encountered the problems that everyone is talking about - the courier got lost, turned pizza into cream soup, ate food along the way.

The plans - an experiment with delivery on scooters by analogy with Yandex.Lavka and Delivery. If it’s not cold in winter, and couriers say no, then scooters are a great thing that solves key problems: time, money and finding people. This will reduce the cost of a kilometer and travel time during rush hours by 3 times, we will be able to attract students to work in our free time. Earn one and a half or two thousand per day of work in Yaroslavl - this is not bad even for an adult man. Scooters can be owned by taxi companies and rented to us. I talked with a couple of taxi parks, they said that from a car bought for 600 thousand - a million rubles, they earn about 600 rubles a day. If you imagine that they will buy scooters worth plus or minus 150 thousand, service them and rent them out to us,we will be able to provide them with revenue of 400 rubles per day per unit. This is beneficial for both taxi companies and us.

Quicq in numbers


To start the experiment in Yaroslavl, we spent 15-20 thousand rubles from our personal funds. This is a bit much for MVP. We went beyond the planned budget of 3 thousand rubles, because we agreed to share all related expenses by all, in a patsansky way. Our costs: thermal bags to couriers (6,000 rubles for 4 pieces), a domain to host a site, small expenses on ads on Avito to search for couriers, 7 thousand rubles for renting an apartment for a colleague who went with me to Yaroslavl.

We started on March 10th and on April 5th we completed 329 orders. Now we have 5 existing customers, 4 in the process of connecting and more than 20 couriers in the database. We earn in the region of 20-30 thousand rubles a week. Of these, about 22 thousand are spent weekly on salaries to couriers, 1000 rubles are expenses on Avito, another 2-3 thousand small expenses. The actual loss that we cover from our pockets today is 6-7 thousand a week.



Quicq through the eyes of IT


The tools that we are currently using within the framework of the project: the site - on the client’s side, the finished B2Field mobile application - on the courier’s side to monitor where it is and in what status. We met with B2Field representatives and asked for a free version for two months. We connect orders with couriers using the B2Field admin panel - this is a flight control panel, its logistics manager, he receives orders, creates routes and appoints couriers. Now the issue of automation is urgent for us, so that the courier is automatically appointed when the order appears.

We understand that in the future we will not be able to compete with strong players in this market using third-party software. Therefore, our primary task is to create our own software with good algorithms, which optimizes the route no worse than others.
We do not have development budgets, we write who can do what, because you have to start somewhere. Fortunately, almost all the guys have an IT background, and it’s very different. In total, we as a team have pretty good skills both in web applications and mobile development, as well as in Machine learning or complex computing systems, such as route optimization. Each team member writes his own block, in this regard, we immediately decided to divide the system into several modules communicating by API.

Quicq is a platform whose core is the main server that implements the business logic of all projects. We initially develop a single personal account for the client, and as new products are launched, we simply raise a new frontend. In this sense, we are inspired by Gojek, an Indonesian company, one of the largest ecosystems in the world. The client has one account from which you can order a taxi, food, call a doctor or dry-cleaner at home, send a parcel and so on.



As a separate unit, we allocated a routing server on which all logistics-related tasks are solved: routes are built and optimized, courier delivery times are predicted, orders are distributed. To build routes, we now use the OpenRouteService engine on OSM maps. Both of these solutions are open source, and we pretty much remade them for ourselves. Long tormented with the choice of cards and routing engine. At first they were skeptical of OSM, but the cards turned out to be surprisingly good, accurate, and in our conditions it is important that they are also free. We probably missed a bit in choosing OpenRouteService, now I would take Valhalla, they are much easier to customize. But quite a lot of its functionality has already been screwed up on ORS, and the main jambs, in general, have been resolved.

All our client applications and internal programs are connected to the main server. For example, we have written a personal account of a logistician / operator, from which you can monitor the movements of couriers and manually manage routes, if necessary. Also, orders from the personal account of restaurants and from the food delivery site fly to this server. On the client side, we only write web versions, this is much faster. For example, to deliver products, we wrote a website from scratch to the first users in two business days. Frontend development time is a very important metric for us, and this is exactly what we are now focusing on the development stack. We write the server on node js, because it is quite fast, plus we have a lot of experience working on the node, this eliminates unnecessary delays. The frontend runs on VueJS because it is corny simpler than React and Angular,and most importantly, it helps to write reactively only part of the site, which means reusing part of the code from old projects.

In the future, we are thinking about integration with iiko or another popular system for restaurants in order to automate the process of creating an application by a client. This is easy to organize, but so far there is no request, because the restaurants themselves are not automated inside. Before our arrival, they did everything in the same way, only worse - they sent information about the order to the SMS messenger.

From day one, we realized that our main goal is to accumulate data. We collect everything that can be collected, track the time of all stages of orders to the minute. Where there is not enough information from the courier application, we put patches in the form of SMS from couriers, which they send to us upon receipt of the order. We put all the data into a database, and then analyze it in Excel. Thanks to the collected data, we created an excellent airbag for subsequent iterations.

The project team


The guys and I met in Skolkovo on the moove program, where we came to pump entrepreneurial and business skills. We teamed up on the principle of "gathering dudes with brains." Our weak point was that we are very similar, we lack non-technical brains. This interferes and helps at the same time. If we were humanities, it would be unlikely that we would fanatically collect data and statistics, but now it seems obvious.

Our team did not have serious conflicts and panic attacks, the work is being carried out as restrained and constructive as possible. We perceive problematic situations as part of the work and do not worry about it.

I spend a lot of time on command synchronization. After each meeting and calling, when we decide how to move on with the project, I write a long, long message, describe what is happening with us, so that it is clear to people out of context, and put it in our general chat. In conditions of self-isolation, when we communicate only on Skype, this helps us all to be on the same page.



Development plans


Yaroslavl is our sandbox, where we want to find the best solution that can be scaled. Now we are writing software and registering IPs, which will give us a 2-3-fold increase in customers. As soon as we understand that our model is working, we will begin to open in 2-3 cities.
We work with restaurants, and several times per hour we receive an application in which there is an address accurate to the apartment, the name and phone number of the person who is using the delivery now. Thus, we get a specific lead, for which we did not pay. What is the point of stopping it?

This week we are launching the regional Yandex.Lavka, a home delivery service in an hour. The competitors we found in Yaroslavl deliver tomorrow and do not accept orders on weekends. The budget at the start is 3 thousand rubles. The action plan is as follows: we take the grocery stores closest to the house, analyze what products are in greatest demand now, parse information on them from online catalogs and upload to our website, made on a free constructor. At the same time, we are thinking about how to organize logistics, we print 300 leaflets and lay them out in houses in the district. If we receive at least 10-20 orders with 300 leaflets, this is an excellent conversion, there is reason to buy traffic for this story and begin to develop. If not, take another hypothesis.

Error handling


Reflecting the experience gained during the launch of this and a couple of previous projects, I made several conclusions that may help you not step on my rake:

  • . , — . , , MVP, . , , — , . , . , . 20 — . , , . — .

  • . , , , . , , . , — / . , , . , .

  • . , , - , . : , , 100 . . , , .

  • = . , , – . , , , . , - , . , .

  • . , “!”. 100% , , . , , .

  • – . , , -. , , , .

  • , , CEO, . , -: , , . , , . , , . : « , - ». . , : «, , - . , ?» , . .

  • , : , . , , - . : « ?» , .

  • , . , , . , - . , « » , , , . , - golama, , , - , .

  • , , . , . , . , .

  • , , , . .

  • , , , , .

More articles:


All Articles