Sizing a £18m opportunity to automate a Big-4 UK supermarket’s clearance process

Turning clearance management from pain point to uplift opportunity

Retailers across the globe understand the importance and pain that comes from stock clearance and inventory management. It’s essential for both financial performance and operational efficiency. It’s often part of a wider scope, taking into account detailed pricing strategies around markdown and assessments of when stock will sell and exit, allowing room in warehouses for new stock to arrive. 

Working with a Big-4 UK supermarket, we saw an opportunity to turn pain points in the stock markdown and clearance process into a revenue generator, making the case for analytics and automation to avoid delays, reduce manual work, and maximise the profitable exit of stock, all within established clearance deadlines.

For QuantSpark, the success of our projects is as much about understanding the existing commercial strategy and business processes in order to make a case for investing in analytics products themselves. This case study shows how we use diagnostics to size the value of clearance opportunity and set out a roadmap for developing a workflow tool that can deliver that prize.

Setting out our approach: Linking analytics projects to company strategy

The focus on using analytics to enhance clearance procedures came from a discussion about the client’s overall approach to product markdown, as part of a company-wide profitability drive.

Their markdown pricing strategy was not centralised, meaning from a financial perspective the discount applied to a particular SKU could vary from store to store, creating conflicts with corporate-level discounting and limiting visibility of the discount’s effectiveness for moving products without unnecessarily sacrificing margin.

Operationally, this resulted in a large build-up of discontinued stock in store warehouses, estimated to be between £10-£15m at any one time, which typically could not be cleared easily to allow for new stock to be brought in.  

The end goal of QuantSpark’s involvement therefore was to diagnose the root cause of this clearance build up and devise a solution, quantifying the estimated return on investment that a proof of concept would deliver. Based on our experience, creating a simple software tool to manage the workflow behind this process whilst visualising key information was likely to be the solution our client needed.

Sizing the opportunity would provide a ready-made business case for our stakeholders to take to the board for further investment in scaling up the analytics solution from Excel-based proof of concept to software MVP to productionised tool.

Sizing the prize: How QuantSpark targets value

Our approach to designing a proof of concept begins with identifying a clear niche within the business where the analytics approach can be trialled, establishing a hypothesis up-front and then set about testing that hypothesis through a combination of qualitative interviews and quantitative analysis.

Selecting a suitable niche is a fast and cost-effective way of demonstrating whether the methodology delivers tangible financial benefit or not, allowing it to be iterated and scaled or wound down depending on the outcome. When selecting that suitable niche, we look for readily available data of good quality and granularity to aid the speed of the project.

For the client, we started with their General Merchandise division before narrowing the focus to Homeware and Furniture. This department often ran into the biggest problems simply from the practical implications of large, unwieldy SKUs. The issue was compounded by the impacts of seasonality on consumer tastes, creating crunch points to move stock in time for the new season, creating a feedback loop where rushed timelines fed increased discounting.

Sizing the prize: Establishing a benchmark

To size the opportunity, we began with the Home & Furniture data that was made available to us, equivalent to the past 400 days of clearance sales. From that period, if the client didn’t apply a discount to their sales, they would have made £300m in revenue. They actually made £250m through discounts, providing a delta of £50m in possible revenue uplift from improving the discounting process – giving our team their benchmark.

Methodology: Combining qualitative with quantitative

Our methodology focussed on:

1. The practical understanding of the clearance problem

2. Modelling & analysis to determine the optimum time to mark down a product during the clearance process.

The first half of our discovery was achieved through qualitative interviews with stakeholders from each business team that came into contact with any stage of the discounting process. Our questions were designed to identify as many problem statements as possible, before using the RICE (Reach, Impact, Confidence, Effort) framework to prioritise those problems.

Reviewing with the client, we picked out three key issues: 

1. Lack of automation: Manual processing during clearance, creating a heightened risk of error which could interfere with timelines

2. Conflict Avoidance: Those delays could be compounded by conflict at a regional and corporate level by driven by poor visibility of the methodology behind discounting

3. Lack of data-driven decision making: Merchandisers discounted on a reactive basis to move stock, instead of proactively assessing where and how discounting would benefit a store’s margins

With the scope clearly established, we worked to size each of the issues, first using surveys to ask how conflict avoidance impacted clearance then applying a weighting to the responses to quantify that impact and find an average.

The lack of data-driven decisions underpinned the second half of the project, using exploratory analysis of the client’s data sets to set out a proactive strategy for optimal discounting.

Identifying margin erosion through analytics

Given the potential revenue uplift of £50m, the key business question underpinning any such improvement to the 8-week markdown system was how to stop the client leaving money on the table through premature or ineffective discounting.

To do this, we picked a single SKU that had gone through the full markdown process and tracked its performance, comparing sales across different stores. We found the strategy could be defined by 4 levers:

• Depth of starting and ending markdown

• Number of markdowns applied

• Duration of each markdown level

• Increment-size of each markdown level

Combining those levels, we could define and standardise the discounting methodology as either aggressive, applying a high starting discount and high increment for example with seasonal items that had a hard deadline for exiting the business, or passive, where discounting more measured.

Outcomes: The case for an automated tool

Operationally, the clearance system featured many manual entry systems that were error-prone and inconsistent, particularly when rushed to exit stock from the business by a required deadline.

The best way to overcome this is through a centralised software tool that could automate these processes and introduce KPI monitoring.

To make a case for investing in the tool, we estimated that just by improving and automating current markdown decision making along the 75th percentile - that is, where we improved 25% of decisions with analytics, we could provide a revenue uplift of 6%, equivalent to £18m. A software tool could incorporate our standardised discounting methodology and allow business users to select different markdown options throughout a product’s lifecycle to visualise the sales impact before applying the discount. It would also feature an alert system to give users price visibility over all SKUs at all times.

This project, and projects like it, are a swift and valuable way of establishing the financial impact automation, analytics and software can have on any business. They deliver clear assessments of potential ROI and a roadmap for creating any software solution.

That roadmap itself is designed along agile principles, testing first with Excel-based proofs of concept before moving onto a software minimum viable product. Each stage allows for iteration and feedback, assessing end user needs and interaction, before creating a fully bespoke and productionised piece of software that can be rolled out across the business.