Source: Martee’s

Martee’s co-founders Harry Slagel and Lucy Adams

Unmanned, self-serve convenience store concept Martee’s has pivoted its business model and is building a demand forecasting software product for the grab-and-go sector.

The company – co-founded by a former rapid grocer Weezy employee and ex-CEO of meal kit brand Plateaway – opened its first unmanned ‘micro market’ at a London co-working space last summer. Several more have popped up in office spaces, hospitals and public spaces since.

While those existing stores will continue to operate, the focus of the new sister business – branded Martee’s AI – will be software as a service.

“We’re not planning to grow the network any further, but the Martee’s stores that we have go on to act as an experimental ‘living lab’ where we can quickly iterate our software, monitor our sales remotely through telemetry and identify the impact of various commercial and environmental drivers through delivering our own proprietary sales data feed,” Lucy Adams, Martee’s AI co-founder, told The Grocer.

To date, the software has been built around Martee’s stores “as our own internal first customer, which has allowed us to move rapidly to get to where we are today” she added.

The company is targeting sandwich shops, sushi stores, petrol station forecourts, and supermarkets with ready-to-eat sections as customers for their tech. It is already working with London-based fresh lunch retailer Abokado with a “pipeline of UK and Europe-based food businesses” launching soon.

The company’s machine learning models combine data from its partner retailer’s point of sale systems with environmental data from third party libraries such as weather, tube strikes and school holidays, as well as inputs “from operatives on the ground” to give the food businesses a more accurate forecast for sales.

“Our models create a richer picture of demand by capturing the factors that affect consumers’ appetite for buy-now-eat-now products in real life; the difference between whether you want a salad or a soup, for example, can vary considerably day-to-day based on whether it’s raining, or maybe rain means you don’t brave going down the street for lunch at all,” Adams explained.

“With growing volatility in consumer footfall due to things like increased industrial action and working from home, as well as having a more variable climate, coupled with growth in online order contribution, it has never been more difficult for a human being or even a simple statistical model to predict demand at all accurately,” she added. “This is why we need AI.” 

Within Martee’s retail business, its use of AI allowed it to reduce food waste by 82%.

“We became really fired up about the impactful commercial opportunity that rolling or demand forecasting AI into a software product for other businesses could represent and, armed with evidence, became increasingly convinced that – despite what seemed like the recent application of AI to just about everything – this concept really did have the legs and customer interest to be a real game-changer,” Adams said.