Feeling good about sentiment analysis

Understanding the role of sentiment analysis in brand safety and suitability

09/18 By IAS Team

This year has been nothing short of newsworthy (dare we say it, unprecedented). With a global pandemic, recession, murder hornets, widespread protest, and a major US election, it’s clear that there’s never been a time where we needed trusted, relevant information more.

But for advertisers, along with the massive boom in content production online comes the age-old question – where should you spend your marketing dollars? And more importantly – how do you know if your dollars are being spent in the right place? 63% of U.S. consumers hold a brand accountable for the content that appears next to its ads – so it’s essential to get placements right.

Safety and suitability are key elements for control over the digital contexts where brands could appear. At IAS, we focus on precise control over adjacencies through a multi-layered approach called Context Control. This suite of safety and suitability products includes a variety of tools for clients to control content adjacencies from standard control through to more brand-specific (custom) control, all without sacrificing scale.

Incremental to protecting brands from appearing next to unsafe content, brand suitability provides more granular control of content adjacencies. Brand safety is for everyone, but brand suitability is tailored to your brand needs.

Keywords alone are not a safety strategy.

With advances in brand safety technology over the past few years, advertisers have tapped into opportunities to protect their brand from negative associations. At the start of it all, the bluntest of tools came into play – creating inclusion or exclusion lists for entire domains or keywords in the respective URLs. But technology has come a long way. For the same reason that these tools were so powerful at the start – their ability to protect brands from large portions of unsavory content – they show their weaknesses as well. “Shoot,” “Shot,” and “Shooting” have topped our customers’ lists of keywords to block for years – but these words can also cause brands to miss out on valuable inventory. For example, beauty brands could easily appear alongside a celebrity’s photo “shoot,” and sporting brands would certainly find an article covering the winning basketball “shot” a suitable ad environment. That’s why you need more than just keywords to find the right environments for your brand.

A little context goes a long way.

Of course, everyone needs brand safety tools to keep them away from risky or unsafe content. But what if they could control where their brand does appear? In the earlier example, a little context goes a long way. Understanding the meaning and intention of the word “shoot” or “shot” ensures that advertisers appear in suitable environments without missing out on valuable inventory and scale.

In fact, the IAS data shows that 76% of content aligned with the keywords “shot”, “shoot,” or “shooting” is actually safe and suitable. That adds up to millions of pages brands could be advertising on, and IAS can help you get there with contextual targeting.

But how does it make you feel?

IAS’s patented semantic technology, a key element to Context Control, uses natural language processing (NLP). This technology dynamically comprehends the nuances of context and sentiment for precise content classification at scale. In other words, this technology analyzes language to understand the context in which it is being used, and assesses the overall sentiment (positive, neutral, or negative) along with the associated emotional classifications (for example, ‘amusement,’ ‘love,’ and ‘hope’ are different types of emotion that can be classified within positive sentiment). IAS research found that content with positive and neutral sentiment tends to create greater engagement: 80% of consumers were receptive to these ads, and 93% more were favorable to the ads and associated brands.

While the effects of positive sentiment are notable, appropriately leveraging the full range of sentiment and emotional classifications can go even further. Sentiment analysis offers a unique opportunity for brands to take control of how they present their message in various contexts. Rather than simply avoiding content, brands can choose how to advertise alongside content of varying sentiments. For example, a campaign message that appears alongside positive sentiment content may differ from the message that same brand wants to deliver when adjacent to other content – this is Context Control for advertising.

It’s important to note that IAS sentiment analysis is not a brand safety feature; it is a suitability and contextual targeting tool that works alongside safety features. Safety features will prevent ads from appearing near sites or content that do not align with brand values or objectives. The sentiment powered contextual targeting of Context Control helps advertisers find content that best matches campaign objectives while safety measures are simultaneously active.

Additionally, a sentiment classification is not a judgment on the value of a given publisher, author, or topic — rather the categorization is simply an analysis of the sentiment and associated emotional classifications of the content on a given webpage. Contextual targeting powered by Context Control is incremental to existing brand safety technologies that are used to avoid content like fake news or unsuitable categories for a given brand like adult content. Content classification helps you find the best contextual adjacencies for the goals at hand.

Check out our Safety & Suitability capabilities now.