There is a lot more to data than numbers, executives from two companies told the INMA World Congress as they wrapped up the Brainsnack session on Sunday.

By sharing a series of theories at the Dallas Morning News, Nicki Purcell explained how she and her team used data to either prove or dispel them. Purcell, chief digital officer at the company, shared a theory that may be widely shared amoung many media companies: If a person is more engaged, they are more like to pay for content.

The team quickly learned that monthly page views and subscription data provided a better understanding of how people became subscribers. To dive deeper into this, Purcell said the team asked another question: “Is it about engagement, or is it about reach?”

By exploring the potential profiles of who their audience may be, Purcell and her team made a surprising discovery: “The reality is that only a subset of the total audience is actually paying us.” 

By examining the data digital subscribers, Purcell’s team realised the majority of new subscribers fall into 50- to 69-year-old category. This worked against the team’s engagement and reach theory.

In testing another theory, the team set out to discover if engaged audiences are more likely to share data that could lead to a new subscription. To test this, the team reviewed past newsletter sign-ups to better understand the conversion rate to paid subscriptions. The data supported that theory, as the team discovered the newsletter sign-ups had a 7.25% conversion rate.

The real lesson in these examples, Purcell said, is about bias. Confirmation bias is a real risk — if you believe that more engaged audiences are more likely to subscribe, you’ll look for data to support that. Along with confirmation bias, she also warned against recency bias, belief bias, and, perhaps the most dangerous, first-person bias.

“I think we’re all guilty in this room in thinking that we are our consumer,” she said.

Arunabh Das Sharma, chief executive officer of Sagacito Technologies in India, continued the audience data conversation: “We tend to leave a lot of money on the table whenever we are doing a transaction. How do you avoid that?”

The answer, Das Sharma said, requires some amount of competition. This is not just about process, he added, but about actually using your data and modeling it to work with current databases.

“The good news is that today with the technologies that are available, it’s very simple, very easy, and very possible to get this done,” he said.

In summary, Das Sharma outlined how the principles are simple:

• You link yourself to databases.
• You link to your transaction databases.  
• You link to some sort of market share so you can look at it competitively.
• Machine learning plays a key role.

In working with Star India, a media conglomerate, Das Sharma said this process involved three steps:

1. “We linked to a lot of live databases that are available in the market but are just not linked.”
2. “We analysed every single transaction, and we started predicting what those transactions might look like in the future.”
3. “We linked it to a system that manages every transaction.”

Das Sharma concluded that the key to data success is the marriage of data and expertise: “A lot of us have the technology. A lot of us have the expertise. It’s about bringing these together with data science, increasing the knowledge database.”

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .