Skip to content

Improving Twitter’s Retention & User Growth Rate

TL;DR ⏩

Problem: Negative Twitter experiences can affect the platform’s retention and user growth rate. I aimed to understand people’s Twitter experiences to improve these metrics. Twitter’s growth rate for 2022 is 2% while its current global retention rate is 84.1%

Role: UX Researcher

Timeline: 3 weeks

Approach: I ran a survey for ten days and explored the data with Excel, Python and a bit of R. 

Insight: I found that 73% of my sample use Twitter for news, 37% for trends and comedy and 28% for networking. Likewise, the quality of Twitter’s content affects how people feel about Twitter while the amount of control they have over Twitter affects how they use the platform. 

Outcomes: I recommended optimising Twitter for news, networking and trends and granting people more flexibility to control their Twitter feeds beyond the available options. If Twitter does this, it’ll improve its global retention and user growth rate.

🔥 I summarised the insights in this case study. Want to read more about what I found and how they invalidated my biases? Read this article.

👩🏾‍💻 If it interests you check out the code I used for the hypotheses tests.

Skip to 💨

My Grouse with Twitter

Twitter is one of the most popular social media networks but it’s not my favourite. When I started using Twitter in 2018, I observed how some Twitter users humiliated others. Twitter harbours noise, hate, insults and unhealthy conversations. In January 2021, I deleted the mobile app drastically reducing my interaction with it. 

A year and a half later, I wondered why irrespective of its toxicity, Twitter was still a social media hotspot. Over 320 million people use it. I wondered if the app’s negativity affected how people used it just like it did to me. 

The Objective

💡 I aimed to understand people’s experiences (what people see, think and feel about Twitter) and how it affects how they use Twitter to improve its retention rate. 

Understanding how people experience Twitter and what affects their use-time is key to increasing its user base. Bad experiences can negatively affect Twitter’s user base and reputation. 

The goals were: 

  • Define how people use Twitter.
  • Discover how people feel about their experiences on Twitter and how it affects their Twitter time. 

To accomplish these goals, I asked the following questions: 

  • What do people use Twitter for? 
  • How do people experience Twitter? 
  • How do these experiences affect Twitter usage?
  • How do people streamline/manage their Twitter experience? Do the current options for streamlining suffice?

Hypotheses

My original assumptions were that Twitter’s toxicity affected the amount of time people spent using it. 

My hypotheses were: 

  1. The proportion of people whose experience with Twitter’s content is awful will be 45% or more of the sample. 
  2. The proportion of people whose feelings about Twitter are extremely unhappy will be 45% or more of the sample. 
  3. The experience people have with Twitter and how people feel about Twitter are significantly associated with how people use Twitter. 
  4. There is a significant association between one’s ability to control their Twitter feed and how one feels about Twitter when using it.

The Approach

I began the project with secondary research. I wanted to understand what had been done, and what insights had been generated. 

💡 I settled on using a survey to answer my research questions because not only did it allow me to measure my insights, it enabled me to numerically establish connections between the attitudes of the Twitter users in my sample and holistically view the Twitter experience. 

Research Process

Recruitment & Planning

Before I shared the survey, I tested it with ten people to ensure that the questions were clear. I considered participants who used Twitter at least once a week. As a screener, I asked, “In the past 7 days, how many times did you use Twitter?” Since I was attempting to understand how people used Twitter and if it affected their use time, it felt suitable to consider those who used Twitter and had some knowledge of the platform. 

Secondary Research

From my secondary research, I found: 

  • Twitter’s retention rate stands at 84.1% globally while its user growth rate worldwide is 2% (in 2022) and is predicted to decline in the coming years. 
  • Gender influences the experiences one has on Twitter: Twitter can be a toxic place for women.
  • Twitter isn’t all that toxic. The experience varies.
  • Twitter has put in place ways for people to customise their experience. Are they enough?
  • People use Twitter because of the gratification they get from the content and interactivity.

I was more interested in the experiences of tweeps, the factors that affected those experiences, and what drove people to use Twitter.

Collecting the Data

I ran the survey for ten days and shared the link on LinkedIn, Twitter 🙈, and various Slack and Discord communities. I received 48 responses; 44 were valid.

Analysis & Synthesis: Invalidating my Twitter Bias

My analysis began by filtering the data in MS Excel. I themed the qualitative responses and conducted a chi-square test of good fit with Python for my first and second hypotheses. For the third and fourth hypotheses, I used a mix of Python and R for Fisher’s exact test since my sample was unsuitable for the chi-square test of independence. I visualised the data with Data Studio.

Insights

🔥 I summarised the insights in this case study 👇🏾. Want to read more about what I found and how they invalidated my biases? Read this article.

👩🏾‍💻 If it interests you check out the code I used for the hypotheses tests.

  • Most people use Twitter as a source of news (73%), comedy (37%) & trends (37%)
Question: What do you use Twitter for?
  • People are in a love-hate relationship with Twitter. These are the features: Positive feelings (55%), News and information (39%), Negative feelings (32%) and Networking (28%).
Question: What do you think about Twitter?
  • Contrary to my 1st and 2nd hypotheses, most people think the content on Twitter is good and feel somewhat happy when using the platform. This might relate to the fact that 73% of the sample uses Twitter for news.
Question: How would you describe the kind of content you see on Twitter?
Question: How do you feel when using Twitter?
  • There’s a chain reaction: Twitter’s content affects how people feel when using Twitter. How people feel when using Twitter affects how they use it!
Question: Does what you feel, think about, and see on Twitter affect how you use it?
  • One’s ability to control their Twitter feed connects to how one feels when using Twitter
Question: How much control do you have over your Twitter feed?
  • Most (39%) tweeps have moderate control over their Twitter feeds because the current options that allow them to control their feeds are insufficient.
Question: How do you manage your Twitter feed to suit your preferences?

Recommendations

  • Give Twitter users more control and flexibility over their feeds. 
  • For those who centre their Twitter use on news, comedy, trends, and networking, optimise the platform to suit these needs. 

Outcomes

Potentially Increase Twitter’s Retention & User Growth Rate

In 2021, Twitter’s global retention rate was 84.1%. The growth rate for 2022 is 2%. What if these rates were more? If Twitter can optimise its platform solely for news, trends, comedy and networking and enhance the ability to control Twitter feeds, Twitter would give its user base maximum satisfaction while simultaneously improving its retention and growth rates. 

Disproved Hypothesis

I must admit, the insights surprised me. The study disproved three of my hypotheses. Only the third one (The experience people have with Twitter and how people feel about Twitter are significantly associated with how people use Twitter) was not disproved. I expected most people to echo my sentiments about Twitter. I was wrong. We can never be too sure of what we’ll find until we conduct research.

Limitations

Small Sample Size

This is the biggest limitation of this study. While my sample size surpasses 30, I still feel my results would have been more accurate with a larger sample. 

Better Survey Questions

Asking unbiased and non-leading questions is difficult. I tried as much as possible to prevent bias in my questions, but I still feel I could have made them better.