SupplyChainTalk: Smart supply chain optimisation - turning trade-offs into payoffs 

Views on news 

Retail sales fell 1.1% month on month in October, the first fall since May, according to official figures from the Office for National Statistics (ONS). Supermarkets, clothing stores and online mail order companies suffered sales declines, which some retailers said was due to consumers delaying purchases in the run-up to the annual Black Friday sales, according to the ONS. Customer demand gets increasingly harder to predict thanks to spikes triggered by social media and quickly shifting consumer behaviour that contradicts historical data.

 

Some bigger players, however, already rely on algorithms that can predict the fall in demand in the weeks preceding Black Friday, while others, especially in the fashion industry, just accept that stockouts and seasonality are an integral part of retail. But accurate demand planning is only half the story as agility can also be boosted via scenario planning and preparation for a number of different future scenarios.  

 

Is end-to-end optimisation easy to achieve? 

The cost versus service trade-off – higher inventory levels will cost more – is a rule that every supply chain expert is familiar with but recently sustainability has also become part of the equation. Sustainability, unless translated into business terms and profitability – is not high up on clients agenda, service providers like HyLogic often find.   Segmentation should be a key element of optimisation in the context of both products and customers. The alignment offered by the S&OP process empowers businesses to pursue the same goal, such as sustainability, across their entire value chain, where they don’t just reduce the plastic content of packaging but also reconsider how shipping mileage can be reduced through better palletisation.

 

There are always trade-offs, but businesses can minimise costs by making considerable gains through sustainability: reducing CO2 emissions is extremely expensive but there are some simple measures that don’t cost that much yet the value they create add up in the long term.  

 

The source of several problems is that there is no consensus regarding how different functions and their heads define supply chains. They usually deal with different aspects of it – order management, production planning or reporting – while these workflows, in fact, are closely intertwined. That’s why it’s key to align company strategy with supply chain strategy. Some may even doubt, though that there’s such a thing as end-to-end optimisation, and when clients come to technologists with a request to optimise, it often turns out that the root cause of a problem is of different nature – not the lack of optimisation but rather functional misalignment or bad data. Coordination shouldn’t only happen between different workflows of the supply chain but their KPIs should be connected as well. Finding the right KPIs to drive optimal operation is also crucial – incentives for procurement, for example, should incorporate the optimisation of inventory levels and discourage overstocking.   

 

The importance of a reliable supplier base is often overlooked. However, with the right contracts, suppliers can also be held to account. For automated decision making, about 99% of the data is already available. It’s planning know-how and a deeper understanding of the business that still fall short. To perfect these, more tradition pattern recognition AI is a perfect tool, while LLMs and gen AI are of little use here. The recommended order for AI deployment is processes first, people second – and only then technology. The best use cases for ML are open-ended problems, where it can support decision making with ideas and guesses.  

 

The panel’s advice 

• Before optimising the supply chain, you must understand its status and inefficiencies.  
• Trade-offs are usually curves – at some points steep, then they level off. Setting inventory levels might cost a reasonable amount up to 96 per cent and then grow exponentially above that.  
• Once you have supply chain visibility, make sure that everyone has access to the view on the same platform and build optimisation on that.  Without seeing the full picture, you may solve isolated problems and throw the whole system out of kilter.  
• End-to-end optimisation works better on the level of strategies. On the tactical level, organisations must have independent optimisations for quick decision making, while keeping their eyes on aligned KPIs.