• by Tobi Elkin  @tobielkin, July 20, 2017

When it comes to addressing brand safety in programmatic, Amit Avner, Founder and CEO of Taykey, a real-time data company that analyzes digital activities, believes that marketers should take a content-first approach vs. a publisher-first approach. Avner shared his views with Digital News Daily.

Digital News Daily: What do you mean by a content-first approach vs. publisher-first approach?

Amit Avner: The home pages of publisher sites have fraud and are a hotbed for bot traffic. There’s a lot of data that indicates people consume and discover content (i.e., through social media) based on relevancy, and not solely based on publishers.

Brands can now buy ads in brand-safe environments in real time without having to guess in advance where relevant content is. If a controversy arises, brands also have the ability to automatically exclude being associated with specific content and publishers, whereas if you’re doing publisher-direct, you have to change buys manually.

Think about it this way: About 30% of web traffic is driven by Facebook alone.That’s not homepage traffic. It’s consumers sharing and jumping from article to video to article. If marketers want to truly get aligned with consumer behavior, they need to start getting more granular.

From a safety and quality standpoint, look at premium news websites. Their brand may be trusted, but much of the content is controversial and not particularly relevant for a slice of the population. What if you could just target the specific articles with positive sentiment that were relevant to your audience across thousands of sites?

DND: What’s your company’s vested interest in this?

Avner: We realized that the real-time interests expressed by the marketer’s target audience are the best way to optimize contextual relevance and inventory quality vs. more static methods like pre-defined whitelists, private marketplaces, and pre-selected keywords and categories.

Our vested interest is in making marketers aware of this opportunity and pushing back on the perception that there needs to be a big compromise between inventory quality and scale in programmatic.

There are hundreds of millions of TV dollars that have yet to move online. Why? It’s still way too hard to reach and engage audiences at scale within quality environments.

After all, there’s not one prime-time experience. People are engaging with a series of topics across multiple platforms throughout the day. Luckily, there’s some consistency to the things that specific audiences engage with, so if you align with the right topics you can effectively piece together a massive amount of reach and engagement. Prime time is a giant puzzle that’s constantly shifting.

DND: What does programmatic direct mean to you? What is the advantage when it comes to brand safety?

Avner: It means using programmatic pipes to buy from specific pre-selected publishers or groups of sites. But this approach, and even the term itself, doesn’t make a ton of sense. The original intent of programmatic buying was to allow marketers to enable real-time discovery of audiences at scale—the "right person, right time, right place" mantra.

Programmatic direct effectively surrenders the last third of this formula right off the bat to someone’s subjective definition of "premium." However programmatic direct does offer a high degree of publisher-level transparency and control. But how helpful are subjective and manual approaches? Increasingly, we’ve seen seemingly benign topics turn toxic at the drop of a hat. Kendall Jenner’s Pepsi campaign is an example.

DND: What do you mean by ‘subjective’ and ‘manual approaches’?

Avner: By subjective and manual approaches, I mean manually declaring certain content "quality" and "safe" at the beginning of a campaign, either via prearranged private marktetplace relationships with specific sites or by manually compiling a whitelist of specific pre-approved websites.

Kendall Jenner has had a series of controversies driving negative sentiment lately, including the huge backlash surrounding her Pepsi commercial.

PewDiePie had a huge scandal surrounding his YouTube channel.

Fox News was swept up in a boycott.
These media and personal brands and the content surrounding them was all carefully vetted and declared "safe," but things can change quickly. Rather than using manual methods, automated sentiment detection can help dodge scandals. It doesn't preclude a blacklisting approach to ensure exclusion of known objectionable content. 

DND: Do you think things have improved in the last few months with respect to brand safety? If so, how?

Avner: I think they have improved from an awareness standpoint, and as they say, knowing is half the battle. This has put some positive pressure on platforms like YouTube to put the full weight of their engineering resources behind the safety problem.