Industry
Retail
Omnichannel CX and Intelligent Innovation – The Future of Retail
Solution
TCS Optumera™
TCS Optumera™: AI-Powered Strategic Intelligence Platform for Retail
Highlights
- Driving the right range into the right location to deliver the highest returns and the best possible customer experience is on every retailer’s wish list.
- However, most assortment optimization strategies take a siloed view resulting in inefficient replenishment and poor availability, impacting sales.
- A data-led approach augmented by AI-powered tools and automation can help deliver coherent layouts and customer-focused ranges that drive sales and profit.
On this page
On this pageinpage
The everlasting bond between assortment and macro space
Store-level assortment optimization without factoring in the macro space delivers sub-optimal results.
Driving the right range into the right location to deliver the highest returns and the best possible customer experience is a win-win. However, in large and complex retail environments, how to go about this is a hot topic. For example, delivering store-level optimization, at the outset, may seem to provide the final outcome in the shortest possible time but will create high levels of inconsistency and drive workload. There is little point in attempting assortment optimization when the macro space allocated to a category is insufficient or over invested. This leads to inefficient replenishment, poor availability, impacted sales, and quite often, poor compliance, as stores change layouts to reflect local requirements, ultimately providing customers with a different experience than intended at a strategic level.
Leveraging AI-powered tools and automation can improve the speed and accuracy of assortment optimization and help identify blind spots created by traditional data analysis. In larger store estates, a multi-step approach to assortment optimization is required to ultimately deliver coherent layouts and customer-focused ranges that also drive sales and profit.
Align space and product assortment with the business strategy for seamless delivery
In retail, the business strategy almost always involves changes to the look and feel of stores, the space product offerings occupy, and the assortments that exist within that space.
Space and product change lifecycles can be long. Aligning activities to deliver business strategy requires careful planning of processes to ensure what was intended is what customers get to see. For this planning to be effective, end-to-end visibility of the product lifecycle and performance reporting against key milestones is required. With so many business functions playing a role in delivering change, a break in the chain of events can have significant downstream impacts. For example, if a retailer plans an above the line (ATL) marketing campaign that coincides with disruptive range and space change activity, customer perception can be damaged, and investment wasted. Assortment and space change should be aligned with broader business initiatives where possible, to deliver unified change and limit the disruption to the in-store experience.
Define inter-departmental layout for the best shopping experience
The first layer of optimization is at the store level and defines space and layout by temperature regime and department.
This store-level optimization involves setting out the overall configuration of the sales floor. As trends and tastes change, store layouts have to be changed to accommodate them. Reviewing layouts with multiple points of view, low levels of data, and competing priorities is highly subjective; a poor decision, if implemented, may take many years to correct. For example, despite the known growth of online retailing for electrical items, some retailers continued to lay down significant space in this area on bespoke equipment. They now find it difficult to cost effectively rebalance the space. With a data-led approach, solid trends can be extrapolated more objectively. AI-powered tools can derive the ideal space needed for the current total assortment and the ideal space based on historic performance at the store level, overlaid with future projections.
Cluster stores to reflect customer needs
Data points on customer types and shopping behaviors are crucial for store clustering.
After deriving the ideal space needed for the current total assortment, the next data points that can be harvested from this algorithmic analysis are customer types and shopping behaviors in groups of stores, which can then be clustered together at the attribute level. However, the number of potential attributes is unlimited, and it is best to classify stores at a high level based on a few key attributes at this stage, where the differentiation in attributes is statistically significant. For example, clustering by affluence and ethnicity (while being aware that generally no location exclusively services a single demographic). This data-driven set of clusters generated by AI-powered optimization tools can then be used to optimize product assortments within the cluster. Clustering allows retailers to drive a sufficiently high level of differentiation through minimal permutations.
Curate the master assortment at the cluster level and optimize macro space
It is vital to review assortment and macro space in tandem for assortment optimization.
After creating the clusters, the next step is to curate the range within the cluster by planogram size. Planogram size is a limitation here, with the existing macro space as a baseline. However, it is important to allow an unconstrained view to output when previous data-led optimizations have not been completed. It may be that for a given cluster or planogram within it, the existing planogram size range is insufficient or over-spaced.
It is vital to review assortment and macro space in tandem because each has an influence on the other; disruptions in macro space are costly to implement and should be addressed with care on a well-managed schedule. Similarly, when making assortment changes, the range churn should be minimized wherever possible, ensuring all new product development adds true value. After attaining higher levels of confidence in the as-is layout, it is possible to independently analyze macro space and assortment. This allows more regular nimble changes to the assortment to keep newness flowing into the market without making more disruptive macro space changes, Such disruptive changes to store space can be managed on a less frequent basis and delivered strategically in line with business strategy, shifting trends, store impacts, refits, and remodels.
Optimize product assortments and automate planogram visualization
The final layer of data-driven assortment optimization is delivered at the store level and can cover many attributes depending on the type of retailer.
Typically, retailers would not want this final optimization to radically change the range, the integrity of the display, or the consistency of look and feel for customers. That’s because retailers are cognizant of the fact that they have already created a clustered range based on the given stores’ attributes. The goal is to take care of the finer changes based on local impacts, micro demographics, and extremes at the store level. Let’s consider a store classified as ‘upmarket’ at a cluster level. Data suggests that sales of champagne and fresh fish are particularly strong and economy brands of fresh bread are driving unacceptable waste. Hence, it makes sense to add more of the former to drive positive outcomes and automatically remove the economy brands of fresh bread. Local products and private labels can also be substituted for national brands as a part of the data-driven assortment optimization process.
Retailers may also have to navigate operational constraints or non-standard fixtures and the assortment can automatically be adjusted to accommodate this. Automating the final step of planogram visualization avoids significant workload while driving a tailored range for the store catchment. Many retailers are making such changes manually, using their discretion based on experience and contextual knowledge. This drives additional rework of plans at the store level or thousands of additional planograms being produced centrally, or a combination of both.
A data-led approach is key
AI-powered assortment optimization backed with a data-led approach can deliver the optimal assortment and macro space mix at the store level and in turn drive sales, margin, and customer satisfaction at a relatively lower overall cost.
To deliver effective AI-powered assortment optimization, it is vital to have strong foundational data (spanning sales, profit, volume, inventory, and customer metrics) and sophisticated algorithms that can not only review historical performance but are also capable of predictive analytics. It is essential to have tools (with intuitive user interfaces) that can integrate with existing systems, ensuring the end-to-end flow of data to avoid manual handoffs between systems that results in low adoption. Finally, you must ensure overall organizational design leverages the power of people process, and technology in unison.
