Adapt or die: how to survive the merchandising evolution
Traditional approaches vs. modern approaches
The entire merchandising process is undergoing an evolution as more roles are taken over by smart algorithms, AI, and platform retailing solutions. Traditional ways of thinking are being replaced with more effective, data-driven merchandising processes that can react in real-time to actual demand.
But what does this mean for the current process of merchandising, and how can merchandising support the sales process with new methods that better serve the customer?
Out with the old, in with the new
The traditional merchandising approach has always relied on a combination of experience, skill, and guesswork. Although this has been the ‘least bad option available, it has always had less than perfect results.
New technology means it is now possible to make better-merchandising decisions (which are backed up with data), and which are more effective at meeting customer demands.
Let’s see how the two approaches differ in the key areas of allocation and seasonality.
The traditional approach tries to guess the number of items customers will want to buy across a long period of time. While this might be reasonably accurate on a macro scale, this approach is hideously ineffective at managing the granular details.
Knowing how many of which shirts are going to sell at each store – for each size – has always been impossible, and it relies on estimating approximate quantities, which are always wrong. Understock is a common result, which means lost sales. It’s an understandable defensive measure, but the lost sales and markdowns are a margin killer. And they’re just unnecessary.
The modern approach uses a smart algorithm that tracks the daily demand for each item and replenishes the stock for each store in small, regular deliveries which exactly match demand. Because we are only forecasting what is needed until the next delivery, there is a much smaller margin of error. Optimal stock allocation is therefore naturally achieved using this method because you will always be sending the appropriate stock to serve the demand in that store.
Traditionally, fashion brands have put out two (or more) major seasons per year. Why? Because the seasons change, the weather changes, and so we need new clothes…ok, sure – but also because it’s good publicity to have two events every year.
The problem is that this doesn’t reflect our modern reality or how consumers shop for fashion. It shows a deeply engrained ‘fashion pusher’ mentality, where the fashion label is dictating what trends will be ‘in’ this season, and then tries to force the consumer to buy those styles.
Shoppers today are more motivated by what is on Instagram today, and not what ‘this season’s’ trend was a month ago.
Gucci, Saint Laurent, Marc Jacobs, and Dries Van Noten all abandoned seasonal shows during 2020 and many brands are looking to permanently reduce or eliminate the number of seasonal collections being released. Gucci is one of those making this change permanent, and will no longer put out seasonal collections at all.
The modern approach doesn’t need to create an entire seasonal collection, because the whole supply model is different. Mass-produced seasonal ranges simply make no sense when you are using a finely tuned algorithm that precisely matches real demand with supply in small, regular batches.
Instead, the data-powered merchandising process will respond to current demand by ensuring a steady flow to the customers. This naturally also requires tweaking of existing supply processes, but this change is already underway as brands seek to forge more resilient and responsive supply chains.
We also don’t need seasonal collections to drum up interest – we have permanent, 24/7 access to our ideal customers via their smartphones and tablets, and we can feed them new styles and trends every day if we want to.
The best part of letting technology ‘in on the game’ is that we can collect and use the data from social interactions and apps to see which consumer trends are emerging right now – and to respond with small-batch productions that field-test and then capitalise on successful trends. Seasonal ranges can therefore be replaced with dozens of ‘hot now’ ranges that capitalize on fresh trends as soon as they begin to emerge.
Data-driven merchandising decisions like these will revolutionise the fashion market, ensuring success for those brands that can keep pace with consumer-driven trends and demands throughout the year.
The role of AI in the merchandising evolution?
Every single process will involve AI at some point in the future, but it will also need a skilled human merchandising professional to instruct, oversee and wield the whole system. While AI is capable of a lot – it is essentially just a tool.
Right now, the most significant results can be achieved when using smart algorithms to determine replenishment and allocation decisions. It can take time to develop custom AI solutions and more advanced capabilities, so a platform is often the best approach to get the ball rolling.
In the very near future, AI will be able to sense trends based on visual information and search data and will be able to auto-suggest trends to experienced human merchandisers and designers who can then start to test the waters with exclusive ‘sneak peek’ 3D-simulated fashion shows that can harvest more detailed data about demand and popularity.
Getting prepared for this new reality will be a chief survival trait for fashion brands in the coming years.
make the transition to a technology-enabled merchandising approach
Mindset – getting granular
- A shift in mentality is needed: instead of focusing on pushing products down to the customer from the top of the chain, we need to change our perspective and focus on which products customers will buy that day in each store – the real demand. When we focus on serving this need our entire retail enterprise becomes a lot more effective.
- It is a mistake to try to determine departmental operational efficiencies without looking at the overall aim of maximising throughput – the algorithms will automatically adjust the flow of goods in the most efficient way for the whole company, so this need not concern us at a departmental level. Using the right technological solution you will always have the right amount of each product, in the right place, at the right time.
- KPIs are a big factor here because they can reinforce departmentalism and entrenched mindsets if they don’t reflect the new goals. Company Throughput should be the chief strategic KPI, and all others should work toward this, such as Understock and Overstock (operational), as well as Inventory/Stockouts and Operational Expenses (tactical).
- A good software solution will provide these KPIs as a part of the package, but these can also be custom developed for your system using your data.
- For your ongoing digital transformation to work, you need the data. This means making structural and procedural changes to ensure that data is collected across the supply chain and especially at the point of sale (which is where we measure demand). Such an update requires both new digital tools and a new mindset.
- Data needs to be collected and accessible to all parties from a central source (cloud-based is more stable), to ensure alignment.
- Start with the immediate implementation of a retailing platform, which integrates with your existing systems to start collecting data without delay. A powerful retailing platform like Retailisation can deliver improvements almost immediately using smart algorithms specifically designed for the fashion industry, and using your real-time data to make data-driven recommendations.