[ad_1]
As the deprecation of third-party cookies looms, contextual targeting is seeing a resurgence. The approach is re-emerging as an environment where publishers can grow audiences and advertisers can maximize engagement to reach the right audiences.
“Contextual targeting goes back to the early days of advertising online — the early 2000s — where it evolved into a form of targeting that used mostly keywords and content on the page without any sort of sophisticated AI or machine learning,” said Lior Charka, vice president, product at Outbrain. “It was a very basic way of understanding what was on the page and then using keywords to target.”
However, as ad tech grew more sophisticated — and complex — and the industry shifted toward behavioral targeting, for many advertisers, contextual fell by the wayside. And with the industry likely to become even more regulated in the near term — contextual targeting is appearing on advertiser’s strategy roadmaps as a future-proof tactic in the face of increased restrictions.
The new era of contextual targeting focuses on sentiment. Marketing teams identify moments throughout the day that tell them about users’ context, such as what devices they are using, whether they are commuting and other elements of their online behavior. To best take advantage of this approach, marketers are building out their first-party data strategies to gather the information they need to inform sentiment analysis and plug into predictive modeling. As they’re gaining traction, marketing teams are unlocking pathways to specific product and activity recommendations, such as the next article to read or area of a site to visit.
Semantic and sentiment analysis are helping advertisers fine-tune placements
Typically, when people refer to contextual targeting, they speak about deep page analysis — what the page or ad is generally about based on understanding the language through categorization. However, in 2023, using semantic analysis, it’s possible to dive even deeper by interpreting the text and deciphering the meaning behind it.
With semantic analysis, teams can identify when and where brands, organizations and people are mentioned online and use natural language processing to recognize how meaning changes when certain words or phrases are present within those mentions. Then, AI processing can recognize the sentiment tied to these signals.
From there, sentiment analysis helps uncover the tone and emotional value of the content. Insights like these can help advertisers who might assume audience cohorts don’t want to appear next to political content identify political articles with a positive tone alongside which they would be comfortable placing ads.
Location and weather data increase relevancy and provide insight into sentiment
When anonymized signals provide the time and location of a page view — and so when and where a user sees an ad — they help advertisers better understand their audience in a privacy-safe manner. With these signals and data, marketers can then infer things about the user, such as what the weather is like, where they are or what relevant e-commerce feeds might be tied to their specific locations.
For example, weather data can increase the relevancy and engagement of ads.
“Weather data can help provide insight into sentiment,” Charka said. “One example is if I’m in London and it’s raining at 8 a.m., maybe this is a great time to show me an ad for a holiday in the Bahamas. But if it’s a Friday at 4 p.m. in July, I’m in a completely different mindset, and I’m looking forward to the weekend, so maybe they’ll show me something completely different. So, weather connects to the location and then to other data signals and brings it all back to the ability to reach the right audience.”
Over the years, a prevalent misconception has been that location data is not privacy compliant. However, privacy compliance is seldom at stake since consent is needed to obtain a user’s location. Furthermore, location data does not need to be pinpoint-accurate to be effective and allow for inferences, so users are additionally removed from the kind of specificity that triggers privacy concerns.
Combining tactics and data points fuels predictive modeling
When advertisers combine sentiment and semantic analysis with location-based signals and context-rich data — especially when leveraging AI to assist with that analysis — they can build user profiles to predict audience behavior. This is known as user intent prediction.
However, a prediction can’t be made based on just one data point (or only a few), so marketers must understand that user intent comes from numerous data points.
These include more traditional information like demographic data, behavioral data and customer feedback, leveraging search queries, clickstream data, social media data and contextual data.
Taken together, these data points fuel relevant content, showing users the next product in which they might be interested. The step is just one in a process that follows; marketing teams must update these analyses over time to identify changes in intent and behavior.
According to Charka, user intent prediction is one result when teams combine all these elements.
“If you’re a big fashion brand and you have all these items in your store, and you’re trying to get people to not just buy something specific that they signed on for, but think about what to serve them next and how to keep them engaged and interested, collecting all of that information and putting that into a predictive model will be beneficial,” said Charka.
When marketers utilize semantic and sentiment analysis alongside location and weather data, one result is increasingly robust user profiles. Another is the ability to predict user intent and behavior. It’s crucial that teams master this tactic ahead of cookie deprecation and increased privacy regulations. When they do so, they are on a path to establishing a future-proof strategy that delivers increased engagement, relevancy and revenue.
Sponsored by: Outbrain
[ad_2]
Source link