Study: Almost half of Singaporean influencers artificially boost followers, comments

Silhouettes of mobile users are seen next to a screen projection of Instagram logo in this picture illustration. — Reuters pic
Silhouettes of mobile users are seen next to a screen projection of Instagram logo in this picture illustration. — Reuters pic

SINGAPORE, Sept 20 — Almost half of influencers in Singapore use artificial methods — including buying followers — to inflate their statistics on Instagram, according to a new report which surveyed 7,107 Singaporean influencers claiming to have between 1,000 and one million followers.

The study was conducted by HypeAuditor, an online Instagram and YouTube audit tool that uses artificial intelligence to detect social media fraud such as artificial followers, likes, comments and other types of engagement. It noted that as at August this year, Instagram had more than 1.97 million users in Singapore.

In response to TODAY’s queries, Alex Frolov, HypeAuditor’s chief executive officer, said that 47 per cent of influencers in Singapore artificially increase their Instagram statistics using various means — sometimes using more than one technique.

He said 27 per cent of influencers here buy followers while one in 10 uses a trick which involves following a massive number of other Instagram users in the hope of being followed back to artificially inflate their social media metrics. Another 17 per cent of influencers here artificially inflate their comments while close to one in 10 uses “comment pods” to do so. That means more than a quarter — 27 per cent — boost comments artificially.

HypeAuditor classifies influencers into the following categories:

— Mega-influencers and celebrities (over one million followers)

— Macro-influencers (100,000 to one million followers)

— Mid-tier influencers (20,000 to 100,000 followers)

— Micro-influencers (5,000 to 20,000 followers)

— Nano-influencers (1,000 and 5,000)

The majority of social media influencers surveyed in Singapore were micro-influencers, with 4,138 in that category.

A total of 96 influencers here claimed to have between 100,000 and one million followers, categorised as macro-influencers by HypeAuditor, while another 999 were mid-tier influencers.

A total of 1,874 were nano-influencers. HypeAuditor said Singapore had some mega-influencers, but they were so few in number — under 1 per cent — that they were excluded from the report.

The highest level of Instagram fraud in Singapore was among the macro-influencers, with 56 per cent of that category cheating on their numbers.

The lowest level of Instagram fraud, at 38 per cent, was among nano-influencers, it found.

HypeAuditor found that influencers in Singapore use the following methods to artificially inflate their growth on Instagram:

— Buying followers

— Buying likes and comments

— Using comment pods — also known as engagement pods or boost groups — which are groups that have been set up by influencers either on Facebook or on messaging platforms such as Telegram with the purpose of collaboratively enhancing engagement on their social media pages

— Using a strategy called “follow/unfollow” where influencers follow other users on Instagram on a massive scale in the hopes that some users will follow them back. A few days later, they then unfollow the users

Popular cheating techniques 

— The “follow/unfollow” strategy was most popular among mid-tier influencers with more than 15 per cent using this method

— Just over one-third of micro-influencers — 34 per cent — bought Instagram followers

— Just over one-third of macro-influencers — 34 per cent — inflated their comments by buying comments or using comment pods

Methodology 

Frolov told TODAY that HypeAuditor uses a machine-learning algorithm to “detect fraudulent activity and low-quality audience”, adding that each metric in the report makes use of a different algorithm to retrieve the data.

The report also used data from a wide variety of sources, including “market research agencies, Internet and social media companies, news media and (HypeAuditor’s) internal analysis”.

“We have collected and aggregated open data from a variety of sources: Social platforms, catalogs, websites, crowdsourcing and many more,” it added.

The algorithm allowed HypeAuditor to divide an influencer’s audience into four categories: Real people, influencers, mass followers and suspicious accounts.

After processing the data, HypeAuditor then “transformed data into intelligent estimations by using best-in-class estimation and machine learning algorithms developed by (its) team of leading data scientists and influencer marketing experts”, the report said. — TODAY

Related Articles