Key Takeaway
If you are the manager of a “food on demand” company or better known as a cloud kitchen then this piece is for you. Read this blog to understand how you can use location analytics to acquire more users, identify bottlenecks in the delivery channel, track your movable assets, and increase profitability as a whole.
Introduction
Food-tech has become an exciting space to be in the last couple of years as an increasing number of players are looking to tap into near-insatiable hunger for the food and beverage market. This has led to the emergence of remote, centralized kitchens.
Every restaurant has a back of the house and a front of the house. Cloud kitchens are nothing but a restaurant with no front house. Cloud kitchens are the next front for the other kind of SaaS business (Space as a Service) and offer a win-win situation for both operators and consumers.
Though cloud kitchens as a concept are new to the world, they are spread across geographies and countries. Some well-known cloud-kitchens are Swiggy Access, Kitopi, DoorDash, etc.
A combination of factors such as growth in smartphones — the increase in delivery services and changes in consumer preferences, have given cloud kitchens a massive boost.
Restaurants vs Cloud Kitchens
The restaurant industry is infamous for its failure rate. The number of independently owned restaurants in the US fell by 2% from the fall of 2016 to fall 2017. The reason could be many but the most common one is location [Source]
Traditional restaurants which used to see barely five-to-ten percent of their total revenue being driven by online orders as recently as three years ago, now see over 30 percent of their revenue coming from delivery platforms.
Cloud kitchens solve this problem by allowing a restaurant to operate without having a physical presence at a central hip location. Consumers get to explore a wide range of choices with delivery services slowly opting to use cloud kitchens to expand their reach. For operators, cloud kitchens are a way to save initial setup costs and property rental costs.
Types of Cloud Kitchens
Let’s look at the types of cloud kitchens and the reason cloud kitchens are gaining popularity in a nutshell:
Hub and Spoke Model: In hub and spoke model, a central kitchen prepares the food and then semi-cooked meals/dishes are shipped to final smaller outlets where they need to be cooked before shipping.
Pod Kitchen: These are small containers that can be placed at any location such as parking lots. Due to their size and mobile nature, they are easy to set up and operate.
Commissary (Aggregator) Kitchen: These are what are popularly known as kitchens as a service (KaaS). These are kitchens that are owned by a 3rd party. Many restaurants use them on a shared basis from kitchen space to fridge space.
Metrics Important for Cloud Kitchens
In this section, we dive into how can the different use cases and metrics relevant to analyze for cloud kitchens.
Demand Side
If you have a mobile app using which your users order food, you would be interested to know the following :
User Acquisition
Acquiring customers or expanding in regions where the demand has already built up is always a good idea since this reduces the CAC (Customer Acquisition Cost). Asking questions like:
Where are the locations where there’s a lot of latent demand already in terms of app installs, searches, and orders places?
can help in understanding which locations are untapped & expanding us cloud-kitchens. In order locations, we can focus on doing very targeted offline marketing or events to create awareness.
User Conversion
When we start looking at user conversion, we basically are talking about churn. While analyzing churn it is equally important to throw light upon “where” churn is happening as it to analyze the “why” whether we have our in-house delivery fleet or using a channel partner or third-party. Asking questions like:
Where do users drop out after opening the app? Where is density of churn high due to servicability? Or in the payment step? Or are we losing out on orders due to ETAs?
can help us identify the supply-demand gaps & bottlenecks in our operations. We can then target some key areas to focus for improving ETAs or serviceability.
If you have a delivery fleet of your own, I highly recommend checking this out:
Towards Data Science
User Activation & Retention
Competing in a market where there are many players and consumers can switch at the tap of a button, user retention is a real battle to be fought for. Asking questions such:
Where do repeatable people come from? What are the order patterns & locations of users who are not repeatable?
To all the repeatable users, we can focus on giving specific promotions to upsell them — either to increase frequency or volume. To users who rarely use our app, we can focus on giving them restaurant-level promotions from their favorite restaurants.
Suppose you find that boxed meals tend to be really popular between 12 PM-3 PM near commercial and residential areas. We can promote bulk ordering or give subscriptions in those areas.
Operations Side
In order to access operations and performance here are some key KPIs which are good to measure.
Improving Efficiency
- Orders Accepted: What is the percentage of orders accepted (Accepted / Received) across all locations? What are the reasons for not accepting?
- Orders Cancellations: What is the percentage of orders canceled?
- Prep time: What is the average prep time across different meals and locations?
- Revenue: Revenue & profitability at each kitchen
- Inventory Management: Which items need to be stocked ? Which items have a higher shelf-life ?
- Menu Optimization: Which items are being ordered by a specific user personas ? Do they have a time trend ? For example, people usually order healthy meals in the beginning of the week while they tend to crave for nachos and burgers during the weekends.
Fleet Performance
- Utilization: How is the fleet utilization spread across the city and how is it changing with the time of the day?
- Deliveries per hour: How many orders can your fleet fulfill in an hour?
- Completion Rate: How many deliveries were unsuccessful because the DE couldn’t reach the right location
Shutting Down
- Demand within 5 km: What is the number of searches within 5kms of the cloud kitchen?
- Order Data: What is the number of orders placed at each location?
- Revenue vs Cost Data: What is the comparison of revenue vs cost data at each location?
New Location Planning
Apart from some internal data of your demand, you can also overlay some external data such as:
- Human Mobility Data: This helps in tracking real-time foot traffic in an area and is useful to pinpoint locations to advertise & expand depending on how frequent it is.
- Economical Data: Demand prediction in space and time (i.e, the purchasing power) can help figure out which food brands should be offered to the population residing in a particular region. This data can also be leveraged for inventory/warehouse planning so as to store only those items which are likely to be procured by the users in a particular region.
- POIs: Important points of interest show malls, metro stations etc along with competitor locations & restaurants. A good strategy would be to locate in a cluster with a number of complimentary restaurant choices and away from any direct competitors.
Build vs Buy: How do you choose?
Locale is a location analytics platform, which can act as a control center ingesting your raw spatial, marketing, demand data to give precise real-time insights about your on-ground assets. In a matter of few clicks, we make all the location data accessible seamlessly in one place, ready for analysis to make precise data-driven decisions.
So the big question remains why should you leverage our platform to track and analyze your ground operations when you could build a tool from scratch for your own need/purpose. Let’s deep dive and discuss on this facet.
Building it Internally
From our experience of working with leading on-demand and micro-mobility companies we observed that most of them leveraged and scaled existing open-source geospatial tools like Kepler or QGIS.
At times developers do hack around and build internal tools for such usage but they aren’t perfectly suited to all business needs and audiences. Also, since these tools aren’t built in a scalable manner hence maintaining them requires a lot of effort and bandwidth.
Why Choose Locale.ai?
So, if you are a company that decides to build this internally, it would have to be built like a platform itself, (much like how Uber has done it). Locale would be cheaper, faster and better for you because:
- Users: A business user or decision-maker who wants the right insights at the right time without depending on engineers or analysts. That’s why our focus was to make it so simple to use so that a user with zero knowledge of SQL can get answers to their questions.
- Scale of data: Our specialty is in handling large scale data across marketing (demand), supply (vehicle, riders), as well as static locations. This creates a single source of all ground analytics for a company and all metrics and knowledge reside here for all teams.
- Both Real-time & Historical Analysis: We also ingest streaming data to give insights in real-time for tactical decisions as well go back in time to do historical analysis for strategic decisions.
- Granular Geospatial Insights: We help you get insights from a city level up to every single order, rider, or journey. we go as granular as a building level.
If you want to delve further, check our website out or get in touch with me on LinkedIn or Twitter.
Similar Reads :
Locale.ai
Locale.ai
