The science behind successful podcasts

Gabriela Huelgas Morales
Posted on Feb 20, 2021

An analysis with applications for podcasters and advertisers 

  Podcast engagement is growing quickly. The podcast U.S audience has grown 20% since 2014, and it is predicted to double by 2023 (The Nielsen Company). Some of the factors driving audience engagement are the use of smart speakers and that audio is the media of choice for consumers during transportation times. Around $1 billion was spent on podcast ads last year, reaching 1.7 million in 2024 (IAB). The increasing business potential of podcasts could benefit from data science. 

    This project aims to help podcasters and advertisers make strategic decisions based on up-to-date data. Podcasters could use this analysis to tailor content to improve audience targeting, while advertisers could adjust their campaigns to attract the right people. To achieve this, I tackled two sub-aims: 

  1. Examine the variables that determine podcast success
  2. Determine if podcast trends correlate with those of Movies/TV shows

      The data I used for this project derives from a global Top 200 chart in which Podcasts are ranked based on audience size. This chart is released weekly, so I used the Wayback Machine internet archive to obtain past charts. I employed Scrapy to extract podcast rankings, ratings, number of reviews, and networks. I ascertained under which genre each podcast is classified from a podcast database (kaggle).

Increase your number of reviews

    First, I investigated if the ratings or number of ratings influenced the popularity of podcasts. I used the rankings as a proxy for podcast popularity. Podcast popularity doesn't increase with the rating (Fig. 1A). In contrast, popularity grows with the amount of ratings only after 100,000 ratings( Fig. 1B). This means that podcasters should incentivize the listener to provide ratings because even swelling the numbers with bad ones will serve to boost the popularity more than fewer good reviews that are all good.

Figure 1

Networks and genres influence podcast success

    Networks influence podcasts in at least two respects. The number of podcasts included in the top 200 chart and the actual rankings of those podcasts. Over the studied time, podcasts produced by TWiT appeared ten times in the global top 200 chart, In contrast, most of the networks had only one podcast in the Chart (Fig. 2A). Similarly, the average ranking of podcasts per network varied widely (Fig. 2B). Clearly, the network is a factor that determines podcast success.

Figure 2

    The genre of a podcast is one of the more relevant factors for advertisers. Specific genres and themes allow for better audience targeting. But, is genre also a determinant factor in podcast success? To answer this question, I studied the number of podcasts included in the chart and their rankings by genre. Culture/Society was the genre with the highest number of podcasts in the chart, followed by Comedy, Business, and Education (Fig. 3A). The average ranking varied widely among genres. The highest rankings were associated with  Film, Music, Comedy, and TV(Fig. 3B). Thus, genre is a factor that has an impact on podcast success. Podcasters could adapt their content, with a few tweaks, to add a specific genre to their classification and boost their popularity.

Figure 3

Some podcast genre trends have a negative correlation with those of Movies and TV series

The Nielsen Company observed: “ as new media options like streaming and podcasts mature, U.S. consumers are opting to add them to their daily routines instead of simply replacing the traditional fare.” The Nielsen Company

    Podcast listeners are likely to watch movies and TV series too. Thus, there is a possibility of podcast genre trends being similar to those of movies and TV series. For example, someone who is interested in true-crime TV series might likewise listen to true-crime podcasts. To explore this possibility, I got the top 100 movies and TV series rankings and their genres. Like  I did for podcasts, I employed Scrapy to obtain the data. I also used the internet archive to get the charts from the same dates as I had for the podcasts ones.

Figure 4

    The trend analysis revealed an unexpected result. Some podcast genre trends had a negative correlation with those of movies and TV series (Fig. 4A). An example of this is the relationship between the trends of Personal/Journal podcasts and those of Horror movies/TV series. Broadly, when Horror movies are popular, Personal podcasts drop in popularity (Fig. 4B). This result has potential implications on the behavior of consumers. If these negative correlations are robust, we could use the movies/TV series trends to predict which podcast genres will rise in popularity. One explanation for the negative correlations is that consumers might be resorting to different media formats to fulfill different needs.

    A perspective on this analysis is to extend the data-gathering period to make it more robust. Hence, I would scrape the podcast and movies/TV series charts for longer in regular interspaced intervals. Also, I would extend the trends correlation to provide data that shall enable podcasts trend predictions based on movies and TV shows.

Data sources

 

check out more or my blog posts here

About Author

Gabriela Huelgas Morales

Gabriela Huelgas Morales

I am a Data Scientist with a Ph.D. in Biomedical Sciences. I enjoy the challenges of solving complex problems, finding meaningful relationships within the data, and providing actionable recommendations and insights. Before joining NYCDSA, I was a scientist...
View all posts by Gabriela Huelgas Morales >

Related Articles

Leave a Comment

No comments found.

View Posts by Categories


Our Recent Popular Posts


View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp