Fishbox is a subscription-based service that delivers fresh seafood packages directly to consumers. Thanks to their rapidly-growing subscriber base, the process of tailoring each Fishbox to suit customers’ stated preferences demanded a complex, mathematics-led solution.
Fishbox was founded in 2014 with the aim of delivering fresh, sustainable seafood from Scotland directly to subscribers across the UK. The popularity of their product range saw Fishbox’s customer base grow from 50 to 1,000 in less than a year, making their original method of manually checking their database and hand-picking the contents of each individual Fishbox unworkable. There were several logistical challenges for them to overcome:
- Customers can choose from three box sizes and state likes and dislikes from a range of 70 seafood products. These preferences must be reflected in their order.
- Fish availability varies both seasonally and daily, so Fishbox staff often have to buy substitute items at market.
- Every fish has a different yield of fillet, so if staff are substituting one fish type with another, they also need to adjust quantities on the fly to avoid shortfalls or excess.
With hundreds of new customers being added, Fishbox really required an automated system that showed them exactly what stock they needed to order at market. Only then could they maintain their values of freshness and sustainability into the future whilst still satisfying the unique wishes of every subscriber.
Fishbox shared their problem with the University of Stirling, and we wanted to help. To kick-start our collaboration, we applied for a Standard Innovation Voucher of £5,000 from the Scottish Funding Council.
Staff in our Computing Science and Mathematics division then set about working on an algorithm that would draw from statistical analysis of Fishbox’s customer database and information on fish availability and cost to create real-time, automatically generated shopping lists for the company. The resulting mathematical framework has dramatically increased the automation in Fishbox’s ordering process without sacrificing any of their trademark personalisation – allowing them to scale more easily and unlock their potential for growth.
Our positive relationship with Fishbox has already paved the way for further collaboration with the company, and the advanced algorithm our team created has potential to solve other real-life business problems.
We helped Fishbox to:
- Create an algorithm that tells them exactly what they need to order to satisfy every aspect of customer demand.
- Fulfil a higher number of orders without sacrificing freshness or customer service.
- Avoid wastage and enhance sustainability by purchasing fish based on precise data, including information about required substitute quantities.
- Become a more scalable business, ready to welcome new customers without fear of how to cater for them.