Solution 13
docx
keyboard_arrow_up
School
Georgia Institute Of Technology *
*We aren’t endorsed by this school
Course
6501
Subject
Industrial Engineering
Date
Feb 20, 2024
Type
docx
Pages
3
Uploaded by prakshipayal
Question 19.1
Describe analytics models and data that could be used to make good recommendations to the retailer. How much shelf space should the company have, to maximize their sales or their profit?
Of course, there are some restrictions – for each product type, the retailer imposed a minimum amount of shelf space required, and a maximum amount that can be devoted; and of course, the physical size of
each store means there’s a total amount of shelf space that has to be used. But the key is the division of
that shelf space among the product types.
For the purposes of this case, I want you to ignore other factors – for example, don’t worry about promotions for certain products, and don’t consider the fact that some companies pay stores to get more shelf space. Just think about the basic question asked by the retailer, and how you could use analytics to address it.
As part of your answer, I’d like you to think about how to measure
the effects. How will you estimate the extra sales the company might get with different amounts of shelf space – and, for that matter, how will you determine whether the effect really exists at all? Maybe the retailer’s hypotheses are not all true – can you use analytics to check?
Think about the problem and your approach. Then talk about it with other learners, and share and combine your ideas. And then, put your approaches up on the discussion forum, and give feedback and suggestions to each other.
You can use the {given, use, to} format to guide the discussions: Given {data}, use {model} to {result}.
One of the key issues in this case will be data – in this case, thinking about the data might be harder than thinking about the models.
Answer:
The homework prompt presents three primary inquiries: A) To optimize sales or profit, what is the ideal amount of shelf space the company should allocate? B) What method will you employ to gauge the potential increase in sales corresponding to varying amounts of shelf space? C) How will you verify the actual existence of the observed effect? To address these questions, we will break down the assignment into distinct components and apply various analytical models at each stage.
Step 1: Eliminate seasonality and random fluctuations to derive the weekly average units of a product sold.
Solution: Utilize Exponential Smoothing on time series data representing the units sold of a product to
eliminate random variance and seasonality from the product's sales volume.
Our ultimate objective is to employ an optimization model to decide on the optimal utilization of shelf space. For retailers operating year-round, certain products like Halloween costumes or greeting cards may experience significant demand spikes in specific seasons. Without employing Exponential Smoothing to mitigate seasonality and random variance in a product's sales volume, the elevated sales figures could bias the eventual optimization model towards allocating shelf space for these products, even during periods when they are not in demand. Consequently, this might lead to missed opportunities to allocate valuable shelf space for more profitable products during those seasons.
Step 2: Enhance the allocation of shelf space for every available product in the store.
Solution: Using Exponentially Smoothed sales volume obtained from step 1, along with product price, area unit per shelf, total shelf space in the store, product name, product surface area per unit, and product profit per unit sold, apply Optimization Models to maximize both the utilized shelf space and the profit.
To address the initial question of "how much shelf space should the company have to maximize profit," employing an Optimization Model stands out as the clear choice. By inputting exponentially smoothed data, unbiased results are obtained, avoiding undue emphasis on seasonal items. The objective function should encompass factors such as the total number of products, units to be sold, and the areas of both products and shelves. Constraints, such as minimum and maximum shelf space for each product within specified thresholds, as well as the overall shelf space utilized, must also be considered. The outcome of this stage will reveal the required units for each product, considering all area restrictions.
Further optimization can be achieved by imposing a constraint, such as using only 90% of the total available shelf space, and repeating the process at 80%, and so forth. A comparison of total sales under different scenarios helps assess the impact of using less than the store's total shelf space. This is crucial information, as if, for instance, sales at 80% of total shelf space equal those at 100%, it indicates that purchasing inventory beyond 80% is inefficient and could lead to lower profits. This insight addresses questions B and C from the primary homework prompt, laying the groundwork for further exploration in
next steps.
Step 3: Recognize products that can be bundled together to enhance sales for both items.
Solution: Utilize K-means clustering on the dataset containing products offered and historically purchased products to identify clusters of high-value paired items that demonstrate mutual benefits from shared shelf space.
Employing a clustering algorithm proves effective in addressing the challenge of identifying pairs of products commonly purchased together and designating them as high value. Specifically, by employing the K-means clustering method, we can establish distance criteria based on whether a certain number of
customers have historically bought the same pair of products. For instance, if a customer has purchased cereal with milk and peanut butter with jelly, they can be grouped in the same cluster as a high-value pair, subsequently allowing for bundling efforts to boost sales.
This approach provides a more detailed response to the second question, "How will you estimate the extra sales the company might get with different amounts of shelf space?" We can refine our
optimization algorithm from step 2 by incorporating the clustered data generated in part 3. As this clustered data reflects diverse shelf space allocations through bundled products, the updated output provides an estimate of the additional sales the company might achieve.
Step 4: Conduct a validation test to determine if the hypothesis of combining products results in enhanced sales.
Solution: Utilize Greedy Multi-Arm Bandit Experimental Design, specifically Bayesian Hypothesis Testing, on the clustered high-value pair items from step 3 and the optimized data from step 2. This approach aims to ascertain the existence of the observed effect and determine whether the hypothesis of combining products actually led to increased sales.
Now that we have the optimized data from step 2 and clusters of high-value pairs from step 3, we can create multiple stocking plans for stores. The initial plan will utilize the straightforward optimization from step 2, followed by subsequent plans incorporating some or all of the high-value pair bundling recommendations derived from our clustering algorithm in step 3. Employing a multi-arm bandit experiment, we can implement different stocking options across various stores and assess the comparative impact on sales to test our hypothesis. This approach offers the advantage of delivering immediate value to customers while continuously gathering information to determine the most effective
stocking strategies.
For this reason, I favor this approach over conducting a simulation where we theoretically compare all options and then select one for implementation, causing a delay in decision-making and potentially missing out on value for the retailer. It is crucial to acknowledge that additional factors such as customer
demographics, store location, and others may come into play when employing a multi-arm bandit approach across multiple locations for a retailer—not solely related to stocking strategies. However, in the context of this case, the instruction is to disregard other factors and concentrate solely on analytics. Given these constraints, I believe this approach would be most effective.
If we reintroduce those other factors, I recommend exercising caution even after gathering data from the multi-arm bandit experiment. Instead of swiftly converting all stores to the stocking plan that appears to perform the best, it's advisable to remain open to adjustments and be cautious about assuming that a plan's success in one store will necessarily replicate at the same level in another.
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
- Access to all documents
- Unlimited textbook solutions
- 24/7 expert homework help