Minnesota-based Fortune 500 Consumer Electronics Retailer needed a streamlined system to understand consumer purchase behavior, consumer reviews, and minimize product returns. The team was spending a substantial amount of time manually analyzing customer comments for returned products and risked missing all the details.
To best categorize the responses into themes and thoroughly understand consumer feedback, they realized it was necessary to invest in Artificial Intelligence. Natural Language Processing (NLP) solution. AIM Consulting was selected as a strategic partner to reduce return rates through NLP, Python Packages, SQL, and Excel Datasets.
Key Results:
- Ability to measure return rates associated with individual products
- Ability to Identify Problem SKUs
- Up to $100k in Savings per Problem SKU
The retail industry continues to be front and center with digital transformation and as such, conventional analytics approaches are quickly falling behind creating a large dichotomy between industry leaders and laggards. AIM provided the necessary industry experience, coupled with our technical knowledge of deep learning techniques, such as Natural Language Processing (NLP), to automate existing manual processes of reading return reasoning and predicting consumer sentiment and behavior patterns to save millions of dollars each year. It’s a really rewarding feeling to see the impactful results and value our solutions provide to our clients, allowing them to be leaders in their industry.
Senior Director, Data & Analytics at AIM Consulting
Project Timeline
2 Weeks
Discover Phase
4 Weeks
Current State Assessment
8 Weeks
Solution Design & Development
4 Weeks
Solution Testing & Iteratitive Development
6 Weeks
Mininum Viable Product (MVP) Deployment
Situation
The organization was using basic summary statistics to determine various bits of information about product returns. However, the team found that the sheer number of consumer data could not be processed manually.
Although catching major themes and problems, the process was still missing considerable data points that would enable them to reduce total product returns. The company wasn’t able to capture product shelf life to add to the manual collection process. As a result, the Client would typically experience a delayed intervention, further increasing return costs.
AIM’s Approach
When working with large amounts of datasets and historical processes, it’s essential to begin a discovery phase that encourages exchanging knowledge between partners. With key learnings from the Returns and Exchanges group, AIM Consulting was able to identify three primary areas in which it could automate processes and data collection.
- First, the team would extract keywords from the customer comments section for returned products to begin to compile enough data to identify trends. This data would provide itself useful.
- The return code was able to be modeled, allowing the team to predict a broader classification of why a product is being returned.
- The team had yet to determine any outlier SKUs that had high return costs. Based on the above return code, AIM Consulting was able to calculate the cost of returns for each SKU and compare it against category averages to determine SKUs generating abnormal return costs.
AIM Consulting was able to leverage a Minimal Viable Product approach that allowed the team to identify meaningful metrics and improvement areas to reduce return rates. Over six months, AIM Consulting utilized Python Packages, SQL, and Excel Datasets to help build out automation that would successfully reduce return rates for the product SKUs, providing roughly $100k in annual savings on an individual product basis—multiplied across thousands of products saving the company millions.
The Results
The NLP solution allowed the queries of consumer product report feedback to be assessed faster and efficiently. Through data processing, excel analysis, and deploying complex python packages for data modeling and visualization, the team can now measure each product’s return rate and accurately understand product sales patterns and consumer purchase behaviors – saving millions in SKU issues and product returns.
