top of page
  • Writer's pictureAdvanced Contextual

Every adtech company says they use AI. But what really matters is what they can actually do with AI that they couldn't do otherwise and how it affects performance for advertisers. 


To understand how AI drives better performance with contextual advertising, let’s review how most companies approach contextual targeting and how AI allows advertisers to generate better performance through more granular targeting.


How most companies do contextual targeting 


Most companies use cookie-cutter taxonomies to facilitate contextual targeting. For example, they might borrow a taxonomy of roughly 300 topics, categorize content on the web based on those topics, and then use keywords to place their ads on content supposedly relevant to their audience.


This is a level of content analysis you don’t need AI to perform because the taxonomy is simple and unchanging, and the targeting mechanism is a simple question of matching advertising keywords to pages of content on which those keywords appear.


The problem with this approach is that it is insufficiently granular, which leads to irrelevant advertising and poor performance. For example, a travel advertiser using keywords to describe travel might end up advertising on a story about people fleeing natural disasters. Or a Bahamas hotel chain might end up advertising on stories that are about travel but have nothing to do with the Bahamas. That doesn’t exactly capture the dream of highly relevant contextual advertising.


Enter AI.


How AI helps Advanced Contextual drive higher performance


AI allows us to process vast swathes of information at a scale and pace impossible for humans. This allows us to do two things.


First, we’ve used AI to develop and leverage a much more granular taxonomy of topics than most cookie-cutter taxonomies. Our taxonomy is 1,500 topics, and we’re able to use it to categorize content more finely, generating more relevant ad targeting and higher performance.


Second, we process 1 million auctions per second. That means we’re able to understand content much better, match the advertiser with a page and audience relevant to them, and drive conversions more efficiently than alternative solutions.


Finally, we use AI to generate synonyms for our advertisers’ keywords. Using generative AI allows us to develop a consistent, rigorous strategy for each client instead of allowing each human team member to freestyle when developing the keywords that will facilitate ad targeting. 


AI is neat. Driving performance is imperative


Every company should be straightforward about where and how they use AI — and also be able to show both how it works and the impact it generates. Let us know if you’d like to see our AI-driven solutions in action. 

23 views0 comments

The end of the third-party cookie on Chrome will make it urgent for advertisers to rethink tried and true strategies such as retargeting. 


Advertisers have been using data from direct publisher buys, data partner usage, and third-party data segments to create site visits for their campaign and retarget consumers with high intent from there. 


But as third-party segments go away, advertisers will lose signal on “intenders,” and parts of their retargeting pools will be eliminated. 


Contextual advertising can help advertisers fill the coming retargeting gap at scale. Here’s how.


Finding your audience more efficiently with contextual advertising


Retargeting efficiencies are directly related to the CPCs paid to create site visits, which themselves create retargeting pools. And the lower the CPC, the better the retargeting efficiency is.


Because context is a good indicator of consumer intent, it can deliver low CPCs that are of high quality. We focus our platform and operations around a few activities to achieve a high-quality, low-cost CPC.


1. We use topics, not keywords. That means that, for a hotel chain in the Bahamas aiming to find audiences who might stay at their hotels, we don't just look for articles that contain the words "Bahamas" or "travel." We identify topics related to your search such as high-end Bahamas hospitality and find a narrower and more relevant set of pages relevant to that topic.


2. We screen out brand-unsafe articles as well as ineffective and low-quality inventory such as MFA sites. You only show up on sites that are brand-suitable and where real humans are consuming content.


3. We only target pages, not domains or channel fronts. For example, the New York Times has pages relevant to a Bahamian hotel chain. But not every page on the NYT is relevant, nor is even every page on its "travel" section. So, we don't target the NYT home page or its travel section. We target the NYT pages most relevant to the topic.


The way most contextual providers would help the hotel chain target potential customers is by targeting keywords. For example, they might help the hospitality business target any articles containing keywords related to "travel" or "the Bahamas." This is how contextual targeting by all the major providers such as Oracle, The Trade Desk, and Google works.


The problem with the usual approach — using keywords alone — is its inefficiency. If you rely only on keywords, you'll target plenty of pages on the Bahamas that have nothing to do with high-end travel. Some may even be brand-unsafe. For example, you wouldn't want to remind prospective travelers to the Bahamas about natural disasters in the region.


This more granular approach to contextual advertising drives much more efficient CPCs. For example, a hotel chain like the example we used in this article used us to achieve $0.80 CPCs against a benchmark of $2. 


With these efficient clicks, you can develop a retargeting pool much more quickly and cost-effectively. Then, you can use alternative IDs to find hand raisers wherever they are until you close the deal.


The end of cookies doesn’t mean the end of retargeting. It just means you need smarter contextual solutions to kickstart the process.

18 views0 comments

The digital advertising industry has recently become hyper-focused on made-for-advertising sites, which waste reach and divert advertiser spend from higher-quality publishers. 


Without a universal standard for defining and removing MFAs, most advertisers are left with either a scalpel or a sledgehammer. Buyers can take DoubleVerify’s “scalpel” approach and screen out sites based on tiers of problematic characteristics. Or they can take Jounce’s “sledgehammer” approach and rely on a cut-and-dry master MFA list. 


But there’s one thing both of these system-level approaches miss. Getting rid of MFAs alone doesn’t guarantee that an advertiser will succeed in a post-cookie world. MFAs are just the tip of the spear when it comes to developing greater precision. 


To target their audience accurately, advertisers need to go beyond ruling out MFAs and screen out all irrelevant content. And to do that, they need to go beyond the system level: they need to dig into each segment, one by one, and rule out any site that doesn’t make sense for a given campaign. 


Advanced Contextual enables advertisers to rule out not only MFAs but also all other irrelevant content. 


Because while some sites may not meet the criteria for being an MFA, they may still score low on attention or be too general to deliver results for a campaign. 


Let’s say a pharmaceutical brand wanted to advertise both a type 1 diabetes medication and a type 2 diabetes medication. Type 1 diabetes may have common ground with type 2 in terms of domains and topics, but there’s a subtle distinction between them. And failing to capture the distinction between the two will lead to underperformance and wasted reach. 


Plus, there may be some sites that system-level MFA audits will fail to catch but that just aren’t relevant to a diabetes medication campaign. For example, a general health website or the Yahoo health channel aren’t MFAs. But if a pharma company trying to promote a diabetes medication advertises on them, the company will reach thousands of irrelevant audience members.


It’s possible to capture nuances like this, but it needs to be done with a precision that working at the system level — or only accounting for MFAs — doesn’t afford. That’s why Advanced Contextual enables advertisers to evaluate all irrelevant content. Because without getting that granular, advertisers won’t be able to screen out low-performing content effectively — let alone reach their ideal customers with precision and efficiency in a cookie-less world. 


MFAs are just the beginning of advertisers’ post-cookie targeting problems. And getting targeting right requires evaluating the proper sites for each individual campaign at the segment level.

12 views0 comments
bottom of page