-
Winning Irish Exporter of the Year, Combilifts CEO Martin McVicar talks about the journey into the Global Market - 19 hours ago
-
JAMES JONES & SONS’ AUSTRALIAN SUBSIDIARY, HYNE GROUP, EXPANDS TO WESTERN AUSTRALIA - October 3, 2024
-
QUECLINK LAUNCHES SOLAR-POWERED ASSET TRACKER FOR TRAILERS AND CONTAINERS - October 3, 2024
-
Southgate Global Calls for Greater Health and Safety in the Warehouse Back Care Awareness Week - October 3, 2024
-
Nulogy to Expand Capabilities with Strategic Acquisition of Mingo Smart Factory - October 3, 2024
-
Cimcorp Group unveils Dreamfield: a new brand to revolutionise the warehouse and distribution industry - October 2, 2024
-
Combilift Retain Deloitte’s Best Managed Company Award for the 12th Consecutive Year! - September 26, 2024
-
Sustainable supply chains are born- how to future-proof your business. - September 25, 2024
-
ZEROMISSION APPOINTS CTO TO DRIVE INNOVATION IN EV FLEET MANAGEMENT PLATFORM - September 23, 2024
-
How to choose a WMS with growing room – six top tips for success - September 20, 2024
Machine Learning.The intelligent way to grow margin
Profits and gross margins are under pressure for fashion retailers. But could Artificial Intelligence hold the key to optimising markdowns and unlocking value in the supply chain? By Dr Manuel Davy, founder and CEO of Vekia.
Fashion retailing is a dynamic, complex and highly competitive business – and it’s not getting any easier. Margins are coming under increasing pressure as a number of factors play-out: Consumers are becoming more demanding, rising inflation coupled with stagnant pay-growth is impacting buying power, costs are going up and online competition continues to gather pace.
A particular challenge for bricks & mortar retailers is the incessant growth of online sales. According to the IMRG Capgemini Online Retail Sales Index for August 2017, clothing sales grew by 17.9%, year on year, and figures from the UK’s Office for National Statistics (ONS) indicate that 15.3% of all retail spending is now conducted online.
With online channels continuing to grow their share of the market, high street stores are under increasing pressure to perform. How can retailers maximise the return from their retail property investments when markdowns appear to be the only way of securing sales and where margins are being persistently eroded?
Sales events are becoming all-important and how they are managed can have a major impact on the bottom line.
Retailers have differing strategies regarding end-of-season sales. Some luxury brands believe they are incompatible with their image, others see them as a necessary evil, but for the vast majority of fashion retailers, ‘Sales’ are now an increasingly important part of their revenue strategy. For them, sales and special promotions are being staged more and more frequently – every few weeks – and big events, such as Black Friday, are gaining huge significance.
Success in maximising the revenue potential of a large estate of high street stores now depends critically on how sales events are orchestrated and managed – and that means accurately determining the optimum percentage markdown on items, the precise geographic allocation of stock to cater for market variances and the meticulous planning and preparation of distribution centres leading into a promotional or sale period.
However, achieving optimum performance across the supply chain in such a dynamic and complex environment is immensely difficult and far outstretches the capabilities of a mere spreadsheet or even reasonably sophisticated sales forecasting methods and tools. It requires a revolutionary new approach.
Big Data, Artificial Intelligence (AI) and Machine Learning, hold the key to unlocking value in the supply chain. By tapping into the vast quantities of data held within a retail organisation, and by combining this with data from external sources, artificial intelligence has the ability to offer potentially transformative insights into future sales performance. Finding the optimum level of discount to effect a sale need no longer be a vague estimate based on a far from accurate sales forecast, and allocating stock to the most appropriate store to achieve a sale does not have to be guess work.
The answer lies in understanding the complex patterns that exist in large pools of data, where the macro can reveal optimal outcomes at a detailed level. Using AI can raise forecast accuracy by over 50% and bring significant financial savings.
According to a report published this year by Ernst & Young, entitled UK Retail Sector: Trading in 2017 – Are you ready for a perfect storm: “Technology and advanced data analytics probably offer the greatest opportunities to improve efficiency. Reducing the amount of discounted stock using more scientific methods to optimise merchandising and mark-down effectiveness is the biggest opportunity to protect gross margin.”
In France, leading retail brands such as Galeries Lafayette and ID Group have successfully deployed AI and machine learning to achieve dramatic results – Galeries Lafayette now has a 97% accurate forecast between projections and true sales figures and ID Group has benefitted from an 8% stock reduction, releasing £55 million of working capital.
Preparing for and managing a sale event is a highly complex activity. A retailer may have hundreds of stores and thousands of skus, which can translate into millions of definable quantities to target.
Planning for a sale starts with a vision of what products are to be on display in a given store during the sale period. The level of markdown needs to be determined and integrated into sales forecasts and the capacity of each shop needs to be factored in. Then there is the planning that needs to take place regarding the capacity and throughput of the distribution centre – how many items can the warehouse manage on a daily basis and how can the capacity be ramped up to optimally meet operational targets?
Determining the optimal level of markdown to achieve sales and minimise wastage is a critical and highly complex calculation. Obviously, a 50% markdown will trigger a greater sales response than a 20% reduction, but what is the optimum level to maximise margin while clearing the stock?
AI and machine learning is used to learn from the past, analysing data on how the sale of a given item has taken place in a given store and how markdown value has affected sales volume. Importantly, this information is aggregated across the whole store geography.
In addition, other factors may be taken into account, such as the weather, school holidays, bank holidays and the day of the week. All this past information is analysed to see which patterns best fit and shape the future sales quantities for each store and each product during the promotional period. Critically, AI successfully looks at large amounts of data, taking a holistic view to reveal hidden patterns and detail that is simply not visible at the item level.
In effect, the amassed data from across the whole geography of the various stores allows you to be more accurate in your predictions for individual locations. This is holistic modelling and it delivers significant benefits.
At Vekia, we have found that by applying our AI generated sales forecast to the replenishment process, leading retail clients have immediately experienced revenue increases of around 4% on average – even though several already had fairly sophisticated systems in place.
Execution is critically important. Leading up to the promotional period meticulous planning is essential, so that the constraints and resources of the distribution centre can be aligned to best effect. Processing and despatching huge volumes of product – hundreds of thousands of pieces a day – must be managed carefully and organised to match the optimal performance of the warehouse as well as the stores’ ability to receive and prepare goods for sale.
In most retail businesses the warehouse is left to plan activity. But by using an AI generated forecast, such as Vekia, to plan and orchestrate picking and despatch operations in the warehouse, optimum performance can be achieved. This has the added benefit of enabling the business to tap into the inherent flexibility of Vekia’s dynamic planning capability, where data can be collected daily, analysed and then acted upon – enabling the business to be far more responsive.
For many retail businesses, organising special promotions and sales activity is a process undertaken largely in the dark. Artificial intelligence, together with machine learning, sheds light on the complexity that surrounds such events and offers the clarity and vision necessary to reduce markdowns and maximise margins. Optimising the retail supply chain takes great intelligence – artificial intelligence.
Manuel Davy is founder and CEO of Vekia, a leader in Artificial Intelligence and Machine Learning for the supply chain. He holds a phd and was a researcher at Cambridge University in Machine Learning. www.vekia.co.uk