Freshlytics

About 2 billion tonnes of food is wasted every year. That much food can feed everyone in hunger and the entire of India as well. Worse, it is equivalent to the carbon footprint of all the cars of Europe across 2 years. 40%of that wastage comes from supermarkets over stocking to tackle the out-of-stock problem (accounting for about 10% of their profit margin). 

We want to help reduce severe food wastage significantly while providing traditional supermarkets with the edge to compete against E-commerce companies such as Amazon. 

Our solution will reduce the food waste generated due to supermarkets across the globe. We strongly believe that if we are able to monitor the food consumption patterns across various supermarkets, we will be better able to assist governments and NGOs with more efficient food distribution strategies.

Globally, it is estimated that over 40% of food waste occurs at production and retail level, with supermarkets accounting for much of it. In Singapore alone, improved food waste management systems can reduce food waste from supermarkets by 400,000 tonnes annually. Unfortunately, current methods to tackle food waste are inadequate and inefficient. These rudimentary methods seek to tackle the after-effects of excess food inventory (e.g. through reduced-to-clear pricing and distribution to food banks), instead of addressing the root cause - inaccurate inventory management practices. Current inventory management practices are still optimised for minimising stock-outs, and supermarkets hold excess inventory at all times. As much as 50% of the perishables in this excess inventory go bad before they are bought, and are thrown out. As a result, food waste continues to be a growing problem in supermarkets and has burgeoned by 40% over the last decade.

Our target segment are medium-large supermarket chains with at least 15 brick and mortar stores, beginning right here in Singapore. We are currently working on a partnership with some of these brand names on a trial basis to better understand their specific needs. 

• Memorandum of Understanding

• Focus group discussions 

• Industry experts and academics as mentors 

While many of these establishments acknowledge the benefits smarter inventory management can bring, creating a separate arm for advanced analytics is simply not in the core business model or bandwidth of these traditional companies. 

Our solution bridges the gap by leveraging on deep learning technology implemented on existing data sources gathered from the system they already have in place. 

These bring 3 key benefits for our supermarket clients to effective tackle their needs:

1. Recommend data-backed purchase and distribution cycles based on existing consumer demand to optimize the shelf life of a product, consequently reducing the need of an oversupply from distributors. 

2. A strong and clear chart for cost savings over time. 

3. Enables them to place more of their time and resources into their core business model- attracting customers to purchase from their stores. 

Freshlytics puts smarter inventory management decisions at the core of supermarkets’ business models. Coupling inventory data with customer transaction data, our advanced data analytics using Artificial Intelligence to provide meaningful insights into inventory flow patterns of each product type. Our solution functions on three levels:

1. Ground employees: Our software provides detailed information on the types and quantities of products that need to be restocked that day. 

2. Store management: Our software provides information on general inventory management based on purchase patterns, consumer demand and turnover rate, relationship with distributors.  

3. Store Chain Management: Deep understanding of purchasing patterns across various stores.

Freshlytics is the first inventory management system to bring data insights down to personnel on the shop floor in an accessible and easily digestible manner, improving their work efficiency. By contrast, inventory management systems have typically only been dealt with at store management levels. In addition, our usage of AI and deep learning allows for continuous improvement of the model, while tailoring each analysis to the specifics of each supermarket store. 

Our solution puts advanced data analytics into the core business model of supermarkets. 

The core technology utilised by our solution leverages on Artificial Intelligence and deep learning technology implemented on existing data sources gathered from their system. We analyse and compare consumption patterns and inventory levels along with other store-specific datasets to continuously refine Freshlytics’ model for each individual store.

Through Freshlytics, supermarkets obtain more accurate inventory management practices (in relation to consumer demand) which are actionable on both managerial and ground levels. Supermarkets no longer have to grossly overstock their products in fear of running out of inventory. This reduces the excess food inventory held at any point in time, reducing the spoilage of perishables in excess inventory and reducing food wastage. 

We are still in the deployment phase. In one year, we will have at least 5 stores fully operational with our software integrated within their workflow.

This year, we will have:

• Completed pilot study with 1 supermarket chain

• Completed Minimum Viable Product with 1 supermarket chain

• Deployed solutions across at least 5 stores 

Within 5 years:

• Expand regionally to nearby South East Asian Markets. 

• Determine scalability of our solution based on analysis of business models, ease of technology implementation, and market segmentation. 

• Continually assess potential of our business to head into other verticals, such as customer profiling

Within 1 year:

• Bureaucracy within supermarket chains

• Lack of standardization of software infrastructure within a supermarket chain

Within 5 years:

• The South East Asian Region has a very low technology penetration rate. This means that many supermarket stores do not necessarily have a software infrastructure in place to track inventory

• There are also numerous small and individually-owned convenience stores/sundry shops in the South East Asian region. These smaller shops might not be able to provide the size of dataset needed to perform our data analytics, limiting the range of our impact. 

• Proliferation of virtual supermarkets in highly developed countries

Technical: Despite the lack of smart manufacturing systems, some supermarket chains (in particular large ones) have digitalised inventory management systems, and we will leverage on the existing datasets. 

Market: Target large supermarket chains

4

The Freshlytics team consists of data scientists, software engineers, economists (specializing in the circular economy), and a business development lead specialising in operations management. Through experience working with local supermarket chains on smaller-scale sustainability initiatives, we have gleaned the necessary industry knowledge for our solution. Our data scientists specialise in deep learning and analytics for complex data sets, and have worked on numerous academic and corporate projects over the last decade. 

We are looking to partner even more supermarkets in Singapore, including NTUC Fairprice, one of the largest supermarket chains in South East Asia. We are looking to analyse their data to provide insights, as part of our pilot study.

Business Model: 

We are primarily a business facing Software as a Service (SaaS) company, with a vision to:

Automate everything in a supermarket to a point where supermarket personnel can solely focus on providing the best customer service. 

In the near term, we are focused on generating valuable insights that help supermarket managers (on the store level and company level) better understand inventory movement and hence, make more accurate and sustainable decisions. In the longer term, our services will extend to software that will: 

1. Seamlessly integrate with suppliers, simplifying the processes associated with inventory management 

2. Provide automatic pricing mechanisms to influence consumer purchasing patterns (in order to reduce wastage of inventory)

3. Reduce human error/work in inventory management

Value to stakeholders: 

Our paying customer are supermarket Chains. They benefit on 3 levels.

1. Personnel: Personnel who were previously focused on inventory management and other laborious tasks can focus on Customer experience

2. Store level: Stores are better equipped to anticipate and handle the out-of-stock problem, while minimising over-stocking.

3. Cluster of stores: With realtime data on inventory movement across stores, supermarket chains are able to benefit from timely information about customers purchasing patterns, inventory movement patterns and associated factors

For a single store with 20 personnel in Singapore alone, we are looking at saving close to USD $1million dollars annually.

Revenue Stream:

We understand that every supermarket store is different so each chain will be paying a one time software development payment between USD $5000 - 10000 and a monthly subscription fees for each store.

Our onboarding strategy includes 2 stages: A trial period accompanied by a paid plan. Our trial period of 3 months will be accompanied by a yearly subscription plan with the supermarket chains. From our experience, supermarkets prefer to have a trial period on 1 -2 stores within the nation, before rolling out across all stores within the chain. At our current stage, we are working with a lot of data dumps, exploring available data to provide meaningful insights on inventory movement. In this stage, we will bootstrap these efforts using grants offered from National University of Singapore Enterprise. Upon gaining sufficient traction, we aim to raise funds from venture capital firms in the region. 

Network:

With the access to a network of entrepreneurs in the Solve Network, we believe we can learn a lot from other entrepreneurs who have worked in a similar domain before. Furthermore, with the ability to share ideas and have discussions with a highly passionate and driven community of entrepreneurs, we believe that we can not only benefit from ideas from across the globe but also contribute to the diversity of ideas within the network of entrepreneurs. More than ideas, a strong network of people is very crucial to fostering a strong support system for us as young founders.

Funding:

With over USD$ 12 million given out in terms of funding, we believe that Solve is  a great avenue for us to raise money to be able to onboard more supermarkets onto our platform and reduce food wastage worldwide.

Mentorship:

The mentors available in the Solve Network would provide valuable industry experience and insights into the supermarket industry, in particular nuances that we should be aware of in dealing with different types of supermarket chains. In addition, we would greatly appreciate business model and scaling advice from mentors with past experience in social entrepreneurship. 

We would like to partner with organisations involved in the logistics and supply chain management space, and food waste management space. 

This can be in the form of accelerators such as Plug and Play, venture capitalists such as Supply Chain Angels (corporate venture arm of YCH Group, Asia Pacific’s leading integrated end-to-end supply chain management and solutions provider) and Sequoia Capital. These partners would provide us with necessary funding and networks to reach out to key industry players. 

We also hope to also partner with established cold chain logistics providers to understand their workings and glean insights on effective inventory practice

With the AI Innovations Prize from Patrick J. McGovern Foundation and Schmidt Futures, we believe that we can use that money and support to develop machine learning algorithms that target reduction in biases and maximisation of data privacy. As we handle data about products moving through supermarkets and tracking of consumer buying patterns, we are very focused on further developing our capabilities to train our deployed models through federated learning and ensure that as much as possible that personal data does not leave the phones of customers and workers. This is of paramount importance to us. Through improving on our existing differential privacy and federated learning techniques that we adopt, we intend to use the cash prize and support to better improve our techniques. This, we believe, will ensure that we respect data privacy laws and also minimise potential biases within the workplace.

Our solution tackles inefficiencies in the supermarket supply chain that results in large amounts of food waste that can be avoided by understanding consumption trends. Food waste contributes to not only a large amount of organic wastage but food production is a large emitter of greenhouse gases. Hence, greater understanding of consumption and potential for optimization of procurement practises can not only provide cost-savings for grocery stores, it can influence upstream decisions and reduce overall greenhouse gas by restricting excessive production. While Singapore continues to resolve downstream food waste problems through introducing composting efforts and using technology such as digesters, Freshlytics hopes to provide a complementary solution by tackling the food waste problem upstream. With a technologically-enabled solution, Freshlytics aligns with Singapore’s national narrative and hopes to help move the needle in co-creating a zero-waste nation.