A few years ago, companies like Picnic realized that the grocery segment was one of the few markets in Europe with a low online penetration and a high revenue potential. Due to the pandemic, market share has increased but so have consumer expectations. An estimated 55% will switch to another retailer if they offer faster delivery; 81% don’t want to pay more than $5 for same-day delivery. Creating a winning proposition against low costs and becoming profitable in the e-grocery delivery market is tough. Shoppers select over 70 items, on average, out of over 10,000 items. To complicate matters, these items can have different temperatures (like frozen, chilled, and ambient) and they can be perishable. Customers need to be at home for the delivery. As a result, almost all e-grocery companies are loss-making, including market leaders like Picnic and Ocado (who do not own physical stores). On top of this, there are challenges in the labor market and high inflation rates.
Overall, the interest in optimizing the last mile in e-grocery can be much stronger, while last-mile fulfillment typically accounts for 25% – 40% of the logistics costs. Every minute costs at least €0.50 in total. Saving a few minutes per order might already make the difference between making a loss and becoming profitable. Let’s explore 4 areas where optimization technology can help.
"Consumers prioritize convenience. They like to choose between various time slots and days, and they prefer short time windows"
Consumers prioritize convenience. They like to choose between various time slots and days, and they prefer short time windows. This creates logistics challenges for e-grocery delivery. Ideally, one van fulfills as many deliveries as possible within a street or area, without visiting it multiple times per day. To achieve this, you can offer only a few time slots per day. Another option is to promote the right time slots by offering them at a lower price or as a greener option.
Steering consumers to the right time slots improves routing efficiency by 10%-15% since zigzagging can be avoided. This saves driving time, giving you the opportunity to accept more orders in the same time slot. This means that time slots can be shown longer on the website, which means more time slots to choose from by the consumer. This is recognized as a service improvement. To assign the right time windows to the right delivery areas and nudge customers in the right direction, you need optimization tooling. This will help you plan routes that are operationally efficient. Alternatively, you can allow customers to choose only one or a selection of limited time slots. This compromises the service level but delivers efficiency gains. If you don't want to compromise on either and offer convenient time slots, use optimization upfront.
"You need optimization technology to run various simulations and select the best scenario."
Optimization can also be valuable to determine your network design and resource mix. To meet service levels, you need to forecast demand per subarea. Next, you need to translate this forecast into a tactical plan and determine the right employee and fleet capacity: how many vans do we need from which type? Do we apply cross-dock hubs and multiple trips? How many drivers per van do we need over the day? Do we want to use electric vans or cargo-bikes to support zero-emission inner cities? Can we balance the workload over the day to create efficient shifts for the drivers? How many employees do we need, and how should our contracts with agencies look like? To answer these questions, you need optimization technology to run various simulations and select the best scenario.
"When you use a cloud-native optimization solution for this purpose, you can save between 5 and 10% in driving time and costs by applying parallel multi-process optimization."
Operationally, routes must be created as efficiently as possible. This starts by proposing the right time slots for each request and steering customers to the right time slot to obtain efficient clustering and routes. When routes are created based on actual and forecasted bookings, high efficiency can be gained. Route calculations require strong mathematical power, given the huge number of bookings and restrictions, like time windows, traffic congestion, allowed vehicle types, possible cross-dock hubs, and capacity constraints. When you use a cloud-native optimization solution for this purpose, you can save between 5 and 10% in driving time and costs by applying parallel multi-process optimization.
"Higher accuracy also translates directly into more efficient plans, which directly creates a win-win situation."
Optimization and AI algorithms are getting much more sophisticated over time. For example, you can learn from historical driving times to provide more accurate estimated times of arrival. Similarly, you can take actual traffic information and weather conditions into account to provide timely updates. Communicating accurate times to the receiver improves service. It also benefits the retailer: stop times are shorter when the customer is ready to receive the goods. Steering drivers with the right navigation instructions and communication improves route reliability during execution, making both your team and your customers happy. Learning from historical stop times also helps you analyze where improvements and higher accuracy are possible. Higher accuracy also translates directly into more efficient plans, which directly creates a win-win situation.
Grocery retailers are looking for opportunities to make their online business profitable. It’s a challenging space with complex business requirements and increasing customer expectations. But there is a positive outlook. For example, Ocado’s e-grocery proposition in the UK is profitable. The company’s overall loss comes from the heavy investments they’ve made in their technology platform. Similarly, Picnic (NL) claims that it can be profitable in the more mature areas, but they have decided to invest in new delivery areas and technology. Esselunga (IT) claims that their e-grocery delivery is now already profitable, and AH (NL) expects to be profitable with their e-grocery proposition in 2025.
Optimization is still underutilized in this space, so there are a lot of opportunities to drive down costs per order. Using optimization technology for last-mile distribution generates substantial positive impact in efficiency, sustainability, and service without large investments. It’s a great time for retailers to embrace optimization tooling and make strides towards profitability.
Sign up to our mailing list and be the first to receive our newest insights and digital magazine in your mailbox.
There has been an exponential growth in online grocery delivery (e-grocery delivery) since COVID-19 hit. At the height of the pandemic in 2020, retailers in the USA saw a 63.9% increase in e-grocery sales. For Europe, the prediction is that e-grocery sales in France and UK accounts for 17% and 22%, respectively, of all grocery sales in 2023. However, grocery retail is a business with small margins. Many retailers in the space are looking to improve efficiency with warehouse automation and robotization. Last-mile delivery is often underestimated. In this article, we share 4 areas where optimization can help retailers drive profitability in e-grocery fulfillment.
By Goos Kant, Professor Logistic Optimization Tilburg University & Managing Partner ORTEC