Is AI the Answer to Profitability in Online Retail?


Single view of stock and an ability to manage and fulfil orders in new ways will be critical to profit for retailers post pandemic, says Yogesh Kulkarni, EVP, COO at AI in pricing, promotion and personalization company,

All over Europe, the proportion of off to online retail has shifted. The average share of online retail sales for UK, Germany, France, Spain, Italy and the Netherlands is projected by CRR to reach 16.2% in 2020, where the average was well under 9% the year before.

This growth will put significant stress on retailers’ ordering and fulfilment systems as they try to manage demand for home delivery, click and collect and an inevitable rise in returns to warehouse and store.

It was inevitable, faced with a unique set of pressures during the 2020 pandemic, that retailers would ramp up their on line operations to cope with a rush of orders, almost regardless of cost. However, this response has eaten into profits for most retailers and is not sustainable.

In order to fulfil this massive rise in demand, retailers had to adopt labour-intensive, short-term solutions in the context of higher labour and delivery costs, and extra sanitisation and safety costs.

If we factor in other elements that add cost – non-availability of goods in an order that will either be fulfilled from elsewhere or lost entirely, new customer cohorts living at or beyond the fringes of traditional catchments, and lower order sizes because the in store impulse purchase is lost on line  – the profit shock hits even harder.

Add the problem of DC capacity and it is not just profitability that is the issue; one US warehouse with a weekly capacity of 500,000 units suddenly saw demand rise to 4 million units. Warehouses also have to contend with illness, industrial action and  stock outs.

Moreover, if we then add to this an ever more demanding consumer, whose tastes and behaviours have changed dramatically during 2020, this simply adds to retailers’ challenges in making predictions for buying, ranging, assortment and allocation.

Finally, add one further variable, proliferation of channels both owned and third party. As social media develops ecommerce capabilities, the question becomes harder to answer – where is the stock, is it optimised by location against demand, can it be fulfilled profitably, is the customer placing the order valuable to me, will running stock down in one location end up disappointing customers?

In this new world, the sheer number of variables involved in optimising both single and cross channel orders is beyond the capability of teams that are generally focused on a single channel, and equipped primarily with spreadsheets that are at best a partial version of the truth and out of date as they emerge from the printer.

The answer is a triumvirate of AI-enabled, integrated solutions that embrace allocation, order brokering and assortment – optimised based on visibility into stock availability across all channels and most profitable fulfilment profile.

Smart allocation maps stores to post codes, based on both store and on line demand (initial orders as well as returns), to optimise supply chain operations, fulfilment and customer service.

Assortment again looks at total stock position to determine the right mix for each channel. And order brokering, sitting on existing distributed order management (DOM) systems, enables retailers to fulfil from stores based on forward-looking weeks of supply and not just the current inventory at a given location.

Profitability is further addressed because this integrated approach can determine the optimal price for items based on demand, availability, fulfilment by channel and sell through.

And because the system enables fulfilment from store based on forward-looking weeks of supply rather than current stock levels at a given location, the order can be fulfilled from the nearest location to the customer, a huge saving in road miles.

As an example, a $2.5bn multi-brand, multi-channel fashion retailer with around 500 stores and 40% Online Revenue wanted to optimise revenue for in-season clearance across all channels, all brands. The company replaced its existing manually intensive tool with Markdown Optimization and Omnichannel Inventory Optimization from Antuit. The results in pilot were an 80% increase in week/week sales and 250% improvement in turns without sacrificing margins.

For retailers that adopt a complete stock management optimisation capability, the gains are impressive.

Better allocation can increase sales by 1-3% of revenue; a price aware ship from store service cuts markdowns by 2-4%; reallocating stock between channels and locations leads to a 3-5 point sell through improvement; while increased store assortments reduce lost sales to recover 3-5% improvement.