Markdown OptimisationKey Benefits
The fashion landscape is constantly changing as new trends arise. Retailers closely monitor these changing trends to aid in purchase decisions and predict sales. As one trend replaces the other, retailers mark down stock to drive the last sales and make space for new items. For many years, the markdown process consisted of applying standardised rules on all sale stock. This method, however, has become outdated as it lacks accuracy and efficiency.
Spatialedge and Pepkor D&A joined efforts to seek a viable solution for this business challenge that retailers face worldwide, and so, the Markdown Tool came into existence.
The Business Challenge
PepAfrica’s standard operating procedures for markdown decisions are set to be applied countrywide per style. Unfortunately, coarsely applying fixed markdowns to the broad store base results in items being marked down in areas that do not require it, pointing to missed opportunities and profits. Another contributing factor that hampers the group’s efforts to optimise their markdown procedures, is the limited markdown budget available to allocate to all the underperforming clothing styles.
PepAfrica realised that it has no means to scalably and accurately identify in which areas markdowns should occur. This makes markdown decisions on a micro level, such as per individual store, near impossible. The group sought an efficient way to accurately analyse data that could be applied to differentiated store- and product-specific markdown strategies.
The combined efforts of Spatialedge and Pepkor D&A led to the development of the markdown tool (MT). The MT is designed to accurately predict sales volumes 10-weeks in advance, offering data to aid decision-makers in their markdown judgments on micro-levels.
Functions and benefits of the markdown tool
The MT can select the optimal markdown, at the right time, for the right level (per region), to ensure that the highest possible profit gets generated. It does so by closely monitoring sales data. If fashion items are selling too slow, the MT will indicate when, and by how much, these items should be marked down. This lowers the usage of the markdown budget while still ensuring that items sell out by the target date. A function like this adds tremendous value to markdown decisions and minimises the risk of implementing markdowns prematurely or too aggressively.
Moreover, to pinpoint where a markdown should occur, the MT can make predictions on various levels. For example, predictions can be viewed by area, store combination, or according to stores with the worst performance. This allows PepAfrica to strategically decide where to apply markdowns as opposed to implementing country-wide markdowns, which are not always as effective or profitable.
The MT further offers additional benefits such as:
Fine-grained view on sale performance and sales predictions
The MT can identify and validate (on the user’s behalf) which clothing styles should be marked down in which areas and by when. This happens on a micro-level (per-store, per-region, per-province) as an alternative to the previous reporting method that only provided decision-makers with a broad overview and summary of the data. The result is faster sell-off rates and, therefore, more clothing styles can be made available in stores. It also improves the allocation of the markdown budget and delivers higher profits due to smarter markdowns, while also marking down clothing styles more efficiently as markdowns occur only in the areas required and not across the entire country.
Automation / Removing human error
PepAfrica analysts typically comb through sets of data pertaining to information of every clothing item in multiple countries every month. The MT can provide a list of slow-selling clothing styles that are subsequently eligible for markdown. By referencing data-based evidence to support judgments, the decision-making process becomes more transparent. Not only does this save decision-makers time and effort, but it also minimises the risk of human errors in the processing of the data.
Combines the best of humans and machines
User biases can unconsciously exclude factors from the decision-making process. By casting all these potential factors into the MT, the tool can take them into account while allowing the user to play with a finite set of parameters if they need to tweak the outcome. This allows the MT to crunch the data, make actual predictions and allow users to provide additional information unavailable to the tool. The MT takes away the emotional component from the user as it presents the user with accurate forecasted data of clothing styles eligible for markdown.
Crafting the Markdown Tool
By applying Spatialedge’s agile approach to software development, we rapidly experimented and generated multiple proofs-of-concept, with frequent user-feedback sessions. Pepkor D&A’s close involvement allowed us to get a good user understanding of the tool. This assisted us in gaining the user’s trust in the tool and accurately delivering what the organisation expected.
Before implementing the MT, the developers wanted to test the validity of its predictions and gain confidence in its results. They did so by testing the tool’s sales forecast on historical data. Results showed the tool’s predictions were at an 80% accuracy level on 55% of the clothing styles tested. According to Malika Naidu, a member of PepAfrica’s central planning team, the company was satisfied with the results as the accuracy level of their current predictions is much lower.
Following this, PepAfrica ran two separate tests.
The first test was a user acceptance test. The user was given a markdown budget and told to implement it according to their current standard procedures. Concurrently, the same test was run through the tool to compare the results.
The second test did not have any user interference. Users were supplied with a list of markdowns generated by the MT and tasked with implementing it.
The table below provides a comparison of the markdown decision process between the user and the MT.
|User (Planner, Analyst or Manager)||Markdown Tool|
|Clothing style(s) chosen||User could force/limit markdowns due to the allocated markdown budget||Picks styles based on logic by calculating the ideal trade-off to maximise the sell off rate|
|Clothing style performance||Based decision on stock volume – favouring bulk over smaller units||Based decision on actual clothing style performance|
|Clothing style timeframe (how long it has been instore)||Do not consider newly stocked clothing styles for markdown||Considers all clothing styles (old and new) for markdown|
|Markdown decisioning factor||How the sales have been performing||How the sales will perform|
|Stock flow rate||High accumulation of smaller units over time as it is excluded from markdown decisions||Higher sell through rate as all units are considered during markdown decisions|
|Risk Factors||Higher risk of human error, selecting the wrong clothing style and time for markdown; human bias||Lower risk compared to users as the tool base its decisions purely on the data analysed and parameters added|
The variances between the user and MT could all be explained and therefore no risk to the business was identified if the choice was made to implement the markdown based on the tool’s results.
Considering this, PepAfrica felt confident to start using the MT.
Results since implementation
Currently, the MT is used in Malawi, classified by PepAfrica as a "one area level"" country. Because of this classification, MT results can be implemented nationally. The decision to first implement the MT in Malawi only was because PepAfrica’s systems are not yet ready to split the markdowns into different area levels. During the May 2021 period, the MT predicted a sell off rate of 94%. An actual sell off rate of 92,7% was achieved after implementing the suggested markdowns across stores in Malawi.
This is an amazing achievement. Accordingly, PepAfrica is focused on rolling out the tool’s predictions into their operational systems. “[Currently] The biggest challenge is to get PepAfrica’s systems to a stage where they can make full use of the tool...” - Elaine De Wet, 2021.
Being able to proactively mark down product styles at an area level within every department in every country will have a significant impact on the business.
The MT is one of many innovative developments that Spatialedge has produced to address the different business challenges of our clients. We offer cutting-edge solutions that are user-orientated and tailored to suit each client’s individual needs.Download PDF