Flawed interpretation, emotional responses 'skew data'
The incorrect interpretation of available data can lead to misunderstandings and misconceptions, and result in bad business decisions, delegates heard at the Informatica World Tour, in Johannesburg, this week.
Keynote speaker Erik Vermeulen, behavioural economics strategist and thought leader, noted that data drawn from emotional responses is seldom accurate. However, careful monitoring of big data gathered objectively could be highly accurate, he said. "We tend to use data to measure behaviour and we tend to forget that it's often the other way around and data influences behaviour. Often we measure the wrong things or our data quality is not right. In the past, we've made a lot of use of focus groups and interviews in measuring consumer behaviour, and it has led to marketing disasters."
He said emotional responses could be influenced by identifying with a particular group or by media reports, citing the example of people fearing sharks when the actual number of fatal shark attacks is far lower than deaths caused by everyday objects, including vending machines. "The leading cause of death in the US is heart disease, with over 611 000 people dying of heart disease annually, yet people in the US appear to fear terrorism more than heart disease," he observed.
He said monitoring factual rather than emotional data could allow businesses to draw more effective strategic insights. "Only by effectively and objectively filtering data, could accurate findings be achieved, he said. "Data filters are powerful - if you know which ones to use," said Vermeulen.
Big data filters monitoring objective data like loyalty card spend were already enabling retailers to effectively track shopping patterns and target consumers with relevant offers. "The careful use of big data and monitoring shopping habits based on loyalty cards can give retailers the edge, as we are seeing in major retailers globally." He cited the example of a teenage shopper who was identified by retailer Target as a possible candidate for baby goods offers, before her family knew she was pregnant.
"So when behaviour is measured completely objectively, results can be staggeringly accurate," he said.