Introduction

As traditional apparel retailers struggle to fill shelves due to the global supply chain disruption, the secondhand apparel market is growing rapidly. Adobe Inc., which tracks online commerce, is predicting a record $207 billion in online spending this holiday season. This year’s secondhand online retail market is expected to top $65 billion, an all time high for the industry according to a survey from Mercari, a secondhand e-commerce site.

In this report, I will use the sentiment analysis tools to perform comparative analysis on TheRealReal and Thredup.

The RealReal is an online and brick-and-mortar marketplace for authenticated luxury consignment. Thredup is a similar online consignment website, but it features a wider a wide range of value, mall and premium brands.

Due to the relatively small number of The RealReal recent tweets, I selected the 100 tweets from each company from November 2nd to November 6th, 2021 to perform the analysis.

Question 1

Which online consignment site has a better reputation on Twitter?

My first question would be helpful to determine which company has a better reputation on Twitter.

My approach is to tokenize the text of the tweets, then assign positivity and negativity using the bing sentiment lexicon. Finally, I summarized the count of positive and negative words for both companies and plot the distribution. It is worth noting since the number of tweets is relatively low, I filtered all words that appeared more than once.

Thredup exceeded The RealReal based on the simple sentiment analysis, with the frequently-used words such as trendy, stylish, clean, sustainability. Thredup also has no negative words that appeared more than once during the timeframe.

On the contrary, The RealReal tweets collected much fewer positive words and more alarming negative words, e.g. trash, ridiculous and worst.

## # A tibble: 1,267 x 3
## # Groups:   query [2]
##    query       word            n
##    <chr>       <chr>       <int>
##  1 thredup     thredup        85
##  2 therealreal therealreal    69
##  3 thredup     https          64
##  4 thredup     t.co           64
##  5 therealreal https          56
##  6 therealreal t.co           56
##  7 thredup     crocs          32
##  8 therealreal leather        22
##  9 therealreal shop           21
## 10 therealreal abel           20
## # ... with 1,257 more rows
## # A tibble: 4 x 3
## # Groups:   query [2]
##   query       sentiment     n
##   <chr>       <chr>     <int>
## 1 thredup     positive     34
## 2 therealreal positive     26
## 3 therealreal negative     22
## 4 thredup     negative     20

Question 2

What are the sentiments people have towards two consignment sites?

In the second part, I am curious to find out the overall sentiments for both companies. Applying the NRC lexicon, I was able to visualize the emotional sentiment counts for each company.

The RealReal scored higher for the majority of negative emotions, such as anger, fear, negative, sadness. By examining individual tweets, I noticed a lot of users expressed frustration over the costly $11.95 shipping fee and its lack of customer service.

Thredup scored higher for most positive emotions, such as joy, positive, and trust. I also saw a number of people using Twitter as a customer service platform to voice their complaints about their Thredup orders, but overall people have decent experience with this site.

Question 3

How does the emotion of tweets change in a day?

I utilized the NRC lexicon to assign emotion to tweets, then calculated the positivity of each tweet by subtracting the count of negative words from the count of positive words. Finally, I used a scatterplot to examine the distribution of positivity during different hours.

For The RealReal, there is a gap from 5 am to 12:30 pm where no valid tweets were captured. There weren’t many tweets during the same timeframe for Thredup either.

The positive positivity scores for both companies were observed from early afternoon (1 pm) to late-night (8 pm).

A number of users attempted to use Twitter as the platform to solve issues with their orders. Both online consignment companies should consider adding customer service specialists during the timeframe when there are more negative tweets, especially late afternoon.