Optimising fulfilment: the backbone of ecommerce logistics

Optimising fulfilment

unpacking the paradox

Changing customer behaviours are reshaping logistics and fulfilment operations as demand for convenient online shopping edges out bricks-and-mortar experiences. Even IKEA, where a visit to a megastore with its play centre and restaurant was akin to a family day out, has reinvented some of its stores to be city-based pop-and-shops, with customers using the premises to check the quality of an item they have already researched and eventually intend to purchase online for home delivery.

As a result of these trends, IKEA is rethinking its business model - and that includes fulfilment, says Olafur Magnusson, their Global Manager of Supply, Fulfilment Planning and Optimisation. With visits to its large out of town stores dropping and online commerce growing, IKEA is looking to use its vast footprint in a different way and the answer is conversion to micro-fulfilment centres.

‘We’re adapting to meet our customers’ needs and creating new types of customer meeting points and using our stores in better way,’ says Magnusson.

This co-location of fulfilment and retail allows the store to take advantage of inventory which is already close to consumers’ homes and, by using AI and machine learning, the company can optimise fulfilment networks across a wide area.

‘We use the technology to connect services and products to provide the best fulfilment set up for each and every customer and in each and every market,’ says Magnusson.

With store space making way for parcel packing, there’s also increased use of technology to accommodate these new workflows. Drones – upskilled with an AI component – work alongside the human to improve the accuracy of stock taking in the warehouses. Advanced automated picking machines can store inventory in the micro-fulfilment centre much more densely and safely than a human agent, maximising vertical storage space, and they can also pick stock for online buyers far more quickly.

While a manual picker does one order with ten items, our advanced automated picking machine can do eight orders with 80 items in the same time.

Olafur Magnusson
Global Manager of Supply, Fulfilment Planning and Optimisation, IKEA

‘While a manual picker does one order with ten items, the machine can do eight orders with 80 items in the same time,’ says Magnusson. ‘The numbers are so good, and if we do not invest in automation we will fall behind.’

He points out this isn’t about replacing people. Instead, the new business model has created a new task – picking – and the company cannot find enough people to fill the roles.

Inventory management has also had an upgrade, with fulfilment companies deploying advanced software that can identify product demand patterns in real-time and optimize the inventory storage strategy to improve efficiency.

Cycle counting, a continuous inventory auditing process to verify the accuracy of inventory records, also gets a reboot with AI, says Martin Seidel, Senior Director, Partner Ecosystem Management & eCommerce Data Analytics at DHL Supply Chain.

‘Using the machine learning, we put in audit data, location data and SKU data, and the model can continuously predict the likelihood of a discrepancy in a given area, which we can then use to prioritise locations,’ says Seidel. ‘With the testing we’ve done, we could double the efficiency of this process, and the error has gone from 5 to 6 per cent to 10 – 12 per cent using the model, which when you operate at large scale is quite a big difference.’

The technology is also helping with returns, which have become a growing financial and stock management issue as e-commerce volumes rise. ‘We have a model that predicts what returns will be, helping our customers make better decisions about how much to order and what can go back on the shelves,’ says Seidel.

AI is also helping companies deal with demand volatility and unexpected disruptions, be they power outages or cyber-attacks.

‘Social commerce is becoming big in terms of demand and there’s real volatility to that,’ says Rejean Kriek, European Sales Operations Manager at Mars, which is using digital twins for scenario-planning and to develop better visibility of risks across the business to build resilience into its network.

The next revolution will be robotisation.

Rob Conijn
Director of Global Projects,
Distribution & Logistics Herbalife

‘Our testing of robotisation has gone very well when it comes to picking stock but the challenge comes with putting it away in the shipper box,’ says Conijn. ‘We want our customers to have a great unboxing experience and for now it’s not there.’

Even in low labour cost countries, we already have quite a significant number of robots working together with humans.

Martin Seidel
Senior Director, Partner Ecosystem Management & eCommerce Data Analytics, DHL Supply Chain

DHL also sees an increasingly robotic future in its warehouses. Its ‘Human Robot Collaboration Index’ shows how many robots work with 100 humans in its operations around the world, and that tally has increased from an average of 53 in 2023 to 62 in 2024.

‘Even in low labour cost countries, we already have quite a significant number of robots working together with humans,’ says Martin Seidel. ‘And this journey will continue over time.’

It’s an early signal of the ongoing revolution in fulfilment. For more insights like this, book your place at Leaders in Logistics 2026 today and be part of the debate.