The increasing use of e-commerce has led to the emergence of many couriers'
service brands. The increased number of courier services has caused fierce
market competition that affects experienced courier brands, POS Indonesia. To
maintain the position and win the competition as one of Indonesia's top courier
service brands, it is important always to evaluate the company's performance. If
the company cannot meet the customers' needs, wants, and expectations, the
percentage of market coverage will decrease, which will result in the company
being unable to compete.
Analyzing feedback/reviews from customers through social media can be utilized
to compare POS Indonesia's performance with other brands, which can later
become valuable insights for improvement. The conventional method such as
interview, survey, a focus group was preferred as a method to collect data in the
past. However, with technological and data developments, text mining methods
can be used to collect data effectively and efficiently.
This study aims to use text mining sentiment analysis to process secondary
unstructured data from Twitter into valuable insights that contributed to the process
to satisfy the customers. The secondary unstructured data was processed through
Knowledge Discovery in Textual Database (KDT), resulting in complex and
uncategorized Voice of Customers (VOCs). The result then transformed into
Critical to Quality (CTQ) Trees to create comprehensive insight that the general
reader can easily understand.
The result of text mining stated that there are five aspects that still unsatisfactory
to the customer: delivery procedure, customer service, agent, tracking system, and
product. The results are already reflecting the real customer feedback that enters
the overall system of POS Indonesia.