Skip to content

Are algorithms manipulating the #healthydiet conversation on Twitter?

Are algorithms manipulating the #healthydiet conversation on Twitter?

Analysis of the 100 most active users in the dataset showed that 81% had a high similarity to bots or had since been suspended from the Twitter platform.

To counter the noise created by bots, marketing automation, and unqualified but highly active and visible users in the healthy diet discourse on Twitter, public health professionals should develop consistent, authoritative, and engaging messaging.

#Healthydiet

A study conducted by The Irish Institute of Digital Business at Dublin City University, in conjunction with safefood, sought to identify the most influential users, communities and topics discussing healthy diets on Twitter. Over 1.2 million tweets featuring the hashtag #healthydiet were analysed from January 2018 to April 2019.

Key findings

  • There is an important distinction between the most visible and most influential users on Twitter. An analysis of the most active users and most visible users suggests that the healthy diet discourse on Twitter is dominated by non-health professionals.
  • Analysis of the 100 most active users in the dataset showed that 81% had a high similarity to bots or had since been suspended from the Twitter platform.
  • As the most active users, low-quality accounts, often bots or backed by marketing automation, are creating noise and confusion within the healthy diet discourse, and may be promoting low-quality diet and health advice.

Digital marketing techniques can be used to combat this noise including:

  • social listening and alerts to identify specific questions or topics
  • scheduling consistent and regular social media posts
  • posting relatable and uncomplicated content to encourage engagement
  • answering questions and including the hashtags and keywords used in the discourse
  • emphasising credentials in bios and posts, where possible

Bots, automation and spam on Twitter

Alongside the rise of social media influencers, discussed in our Influencer bulletin, recent years have seen an increase in the use of digital tools on social media platforms to increase followers, brand awareness and influence. These tools include: 

  • targeted advertising
  • marketing automation and message scheduling
  • bots and spamming
  • bought followers

Key to a better understanding of the healthy diet discourse on Twitter is an understanding of the usage of these new software- or algorithmic-based tools.

The study highlighted that marketing automation, and specifically bots, on social media can come in several forms. These can be used for benign purposes, for example, activity tracking, marketing productivity and customer service. Unfortunately, they can also be used maliciously - for  spamming and manipulative marketing.

Activity, visibility and bot usage

To put use of bots in context, the study examined the difference between the most active and visible users within the #healthydiet discourse.

The study defines the most active users as accounts which generate the greatest volume of tweets, retweets and replies, while the most visible users being accounts with the greatest volume of retweets and replies.

To understand the difference between these users, and to identify possible bots within the dataset, the study used the IUNI Botometer. This is a machine learning algorithm used for detecting social bots on Twitter. It leverages features from a Twitter account and its activity to evaluate the similarity of that account to the known features of social bots. These include user-based, network, content, language, and sentiment features.

This analysis found that 60% of the 100 most active users in the #healthydiet discourse had high similarity to bots and 21% had since been suspended from the platform. This indicates that low quality accounts or bots account for 81% of the most active users and as such are contribute a significant portion of the messaging within the #healthydiet discourse.

In contrast, highly visible users had much higher levels of legitimacy and considerably lower similarity to bots. Unfortunately, they were also significantly less active.

What should public health professionals do?

The study shows that the healthy diet discourse on Twitter includes the use of sophisticated marketing automation tools and the widespread use of bots. These create a lot of noise and potential confusion in the discourse on and around healthy diets.

However, these tools can also be used by public health professionals to amplify authentic, expert and authoritative messaging. These tools include:

  • marketing automation to schedule regular posts
  • social listening for alerts on key topics
  • highlighting consistent and relatable content to engage and inform users

Using these tools, public health professionals and bodies can develop effective counter messaging as well as increase their influence by being both visible and active. To do this effectively, professionals/bodies are recommended to follow an iterative approach of monitoring, analysis, planning, execution and learning, through controlled experimentation.


Read the full study: Sorting the Healthy Diet Signal from the Social Media Expert Noise: Preliminary Evidence from the Healthy Diet Discourse on Twitter

Read further bulletins from the study on Influencers and Topics in the #healthydiets discourse.


Webinar: Exploring the unqualified expert community online, Prof. Theo Lynn, DCU



Safefood Logo

Sign up for our latest healthy eating and recipe updates.

Safefood logo

The site content is redirecting to the NI version.

Confirm