A novel packaging evaluation method using sentiment analysis of customer reviews

Abstract

Packaging has an important role in protecting products during the distribution process. Therefore, it is essential for the packaging designer to evaluate the packaging even after the design implementation. While every new packaging design needs to go through a series of physical tests to ensure it delivers the product to the customer safely, it is not uncommon to encounter packaging failures after the design implementation. Redesigning a completely new package and redoing these tests is very expensive and a time‐consuming process. Therefore, an alternative evaluation method is needed to enhance efficiency while maintaining evaluation quality. The main purpose of this paper is the evaluation of package performance during distribution using customer reviews. Sentiment analysis (SA) is implemented to identify the positive or negative sentiment from customer reviews. Moreover, the proposed method utilizes an in‐house library containing packaging related words (Pack‐List) to identify the reviews concerning the packaging. As a result, this method provides a systematic approach to detect problems by analysing customer reviews as feedback to a defective packaging product. By using the results of SA with Pack‐List, a percentage of negative and positive reviews are calculated to examine the packaging performance during distribution. Then, the percentage of failure over various months and years is examined to identify how time affects the packaging failure. Next, a word cloud of negative sentences is created to show the most mentioned issue. With the proposed method, packaging designers can identify packaging problems in the early stages and keep track of the packaging performance.