Audience data is undeniably an important ingredient in any ad campaign, but some marketers are misusing such data to achieve easy success.

ExchangeWire spoke with Bryan Melmed, global vice president of insights at Exponential Interactive, who championed the use of data intelligence and analytics to support advertising efforts and ad targeting campaigns.

But how are marketers wrongly using data? According to Melmed, brands are missing the point when it comes to understanding their baseline or the audience they should be focusing on.

"Marketers will advertise to consumers who are already passionate about and loyal to the brand. So the measurement will inevitably show their campaigns are successful because they're targeting people who would already go to their website and make the purchase," he explained. "It's like a self-fulfilling prophecy."

More brands recently have been resorting to such audience targeting initiatives, especially in Australia and Southeast Asia, where they are "so desperate" to run successful campaigns that they are choosing to preach to the choir. They end up targeting audiences they know already visit their website regularly and have purchased one of their products or service.

Bryan Melmed, Exponential Interactive

"It's difficult to measure success in marketing and if you take shortcuts, it looks like you're doing the right thing because you've put blinders on," he said. "To really grow your audience, it takes more courage and money because you won't always win. They're harder to find and it's harder to communicate with people who aren't naturally receptive. You want to set aside the people you know are likely to be your customers and those who are not, and look for the middle ground."

With brand safety a persistent issue in programmatic buying, data also can play a role in addressing this challenge, but it alone will not be sufficient.

Checks and balances still needed in programmatic

Melmed noted that computers are intelligent and can identify the cheapest media, and at the right purchase price, if instructed to do so. It boils down to the human to set the required rules and "incentives".

"You can say to a programmatic buying system it can't do this or that, but there will be that one thing you forgot to address or didn't know was out there or there can be fraudulent sites. The computer will sense an opportunity and in most cases, you get what you're paying for," he said, suggesting that premium content typically comes with higher buying price.

"It's an environment that needs to be managed. If you make it a game where loopholes are always the winners, then the system will find the loopholes," he added.

Melmed also noted that while most commodity industries had an authority to manage and ensure things were working well, this was not the case for the programmatic market.

He explained that where auctions were concerned, it would be generally true that the winners were also the losers because they had overvalued the item that was up for sale.

He further pointed out an association with the 'lemon law', where the seller typically had more visibility and more information than the buyer. He referred to buying a used car as an example, where it would be impossible to tell the actual condition of the car unless the seller truthfully divulged all information about the vehicle. This scenario played out the same way in the programmatic media industry.

"The industry is chasing its tail, so to speak. It's looking for blacklisted publishers and more visibility, but it still needs to implement the necessary checks," he noted. "It's like putting all the antivirus software you want in your computer, but you still need to be careful which websites to visit and attachments to open. It's the same for programmatic."

"Data in this case would be like the antivirus. You can set up the rules and data can help cut out 90% or 95% of the fraud, but some computers on the other end would try to evade these rules. If you look at the research on ad fraud, it'll tell you that computers can now mimic a human's browsing pattern so it's almost impossible to tell one from the other."

While these challenges were unfortunate, Melmed said they should not stop brands from adopting programmatic buying because it offers several benefits. Instead, industry players and brands should work towards establishing relationships based on trust.

Data volumes could further increase with the adoption of wearable technology and the Internet of Things (IoT), where web-connected home appliances such as refridgerators and ovens as well as cars would contain information about consumers.

Melmed, though, noted it was unlikely information collected from such devices would be useful for marketers. "When you start to measure things that are particular to a person, such as whether they have exercised enough for the day, that's not something we want to know because it's too private, and for us, too uncomfortable to collect."

"Digital marketing is about volume and understanding the crowd. We want to look at information that people want to share with others, not how well they sleep at night. It's so personal that it's no longer relevant to advertising," he said. "In a way, that's too much data and I don't see IoT data as useful to marketers anytime soon."