Unruly claims algorithm predicts video virality

The launch comes three months after the company opened its regional headquarters in Singapore.

Unruly said it has tuned the ShareRank algorithm, which it launched in the US and UK in 2013, specifically for countries across Southeast Asia. The firm, which has distributed branded content from Unilever, Kia and LG, claims that advertisers using the predictive technology will be able to identify the psychological, social and creative triggers that drive the success of their video content in various territories. Ultimately, this will reveal the potential of earned media on a video ad before they have spent a penny on media.

According to APAC MD Phil Townend, demand for Unruly's full suite of services, including social-content strategy, benchmarking, analytics and distribution, has grown tremendously. 

When it comes to predicting video success around the world, there’s no such thing as a one-size-fits-all approach, he said. “The same content can perform very differently depending on the launch market and demographic in one market versus another.”

As such, the company tests regional versions of the ShareRank algorithm against target demographics to ensure its predictions account for cultural differences. When adapting the algorithm for Southeast Asia, Unruly found that audiences in the region have stronger emotional reactions to video content than any other part of the world.

The firm discovered that the intensity with which viewers feel primal responses (fear, anger and arousal) when watching a video better predicts sharing in ASEAN regions than other markets. For example, in Singapore, a viewer feeling intense anger when watching a video will be more likely to share it than an angry viewer in the UK.

The company also discovered that average share rates for branded videos in Southeast Asia stood 2.44 per cent higher than the global average of 1.31 per cent.

Phu Truong, MD of video ad platform company TubeMogul  said that while tools like this are great for the industry, it is virtually impossible to predict virality. “Virality is about emotion and it’s really hard to calculate how people will respond.”

Jones and team constantly train the algorithm with new data. At present, the tool combines 430 billion video streams with more than 100,000 consumer data points. Unruly records information using self-reported consumer panels, in which respondents are asked to report the intensity of psychological responses they feel on a scale of one to 10. It measures the strength of viewers’ intent to share videos and their motivations for doing so.

“It is very important when brands create an ad to give viewers a reason to share their video," Jones said. "It’s one of the biggest variables that affect an ad’s share rate."

Examples of such social motivations are shared passion, social good and reaction seeking.