AI promises lower costs in last mile delivery

Stuart is an odd name for a company, which is perhaps appropriate as this is a very different sort of logistics firm. It is not focused on hard assets such as depots and directly employed drivers, vans and forklifts. Rather, its value derives from an algorithm that seeks to efficiently match up retailers and consumers with a network of self employed couriers for last mile deliveries.

Traditional last mile logistics uses a routing model, with the courier operating warehouses, and dispatching vans with its clients’ packages.

In contrast, Stuart uses what it describes as a density model for its fleet. It is a model which Accenture predicts could dramatically lower the costs for last mile deliveries. 

Stuart’s UK MD David Saenz explains: “We plan nothing in advance and rely on our forecasting of client demand in order to project how many deliveries clients will need in a given day, in a given zone. This allows us to do flexible pick ups and faster drop offs. We can use the clients’ retail locations as our pseudo distribution centres.

“It means instead of delivering the package from 100 miles away, they can do so from one mile away, for example. You get the product to the consumer much faster, and it is more amenable to the consumer changing their mind. If a product is on a truck using a routing model, you can’t change tack, the truck has to go through the schedule.”

Stuart uses the algorithm to forecast demand and take account of other variables such as weather, events, clients changing mind (how often that occurs and what lag should build into the system to account for it). The software assigns variables, and coefficients of importance to these variables. These are applied to order requests, and the algorithm decides the best courier, route, pick-up point etc.

“We have some very smart guys with PHDs working on the algorithm, but the more variables you add, the more complex it becomes. When you jump from 10 to 50 variables, the complexity increases by 500 times. This is a long term investment, we have a very good version today, but in 12 months we will need it to be exponentially better.”

Stuart is a young company that launched in Paris in December 2015, and now covers a number of cities in France and Spain, as well as Zones 1-2 in London. It plans to launch in a further 3-5 cities in the UK by the summer. Retail clients signed up so far include Sainsburys’ and Kiko Cosmetics.

David also sees the Stuart system as retailer friendly because it does not hold or use customer data.

“We are not a marketplace like Deliveroo or Amazon. The retailer owns the customer data and we don’t have any interaction other than handing the package over. This resonates very strongly with retailers.”

Retailers interact with Stuart in one of two ways, either via API integration at point of sale; or via Stuart’s web solution which asks consumers for pick up and drop off point, package size, etc (aimed at smaller retailers).

Using the self-employed contractor model is potentially controversial – think the DeleteUber campaign – but David says Stuart is working hard with its couriers.
“We ensure couriers earn well above the London Living Wage, we involve them in the decision making process, and strive to make them fell part of a community.

"They are genuinely self employed and do not have to be available exclusively to Stuart.

“We also train couriers when they are are onboarded, we have numerous requirements from retailers emphasising courtesy, friendliness and so on, which are very important also.”

I asked David if logistics was unusual for a start-up company given the traditionally low margin nature of the business.

“There is money in this, we don’t deal with moving pallets around,” he said. “Our economies of scale come from density, we can see the numbers we need to hit, in terms of multi drops, etc. We need a healthy number of pick up points and lots of orders per hour to keep couriers active. The better we achieve this, the more attractive the business becomes from a margin perspective.”