Why AI-powered targeting matters in contextual advertising

Contextual advertising is a hot topic, but has it advanced enough to give the appropriate alignment and relevancy for your advertising?

Key points:

• Contextual targeting is a method utilised to place your adverts within a relevant and appropriate “environment”, but a good environment could simply be the homepage of a targeted website. We therefore that contextual targeting should only focus on relevant content.
• Based on research, we estimate that 20% to 30% of ad impressions served via contextual providers are simply served on the homepage of a site, or just a section of a site. 
• Alphix Media Exchange’s deep contextual targeting tool can fully understand the content itself, identify synonyms, context and semantically match it against our search phrases. As we scan across millions of articles, it allows us to advertise against only the most appropriate pages. 

Contextual targeting is in a period of real resurgence, driven by factors such as cookie deprecation and e-privacy regulation. It is now an estimated $261 billion industry and will continue to grow, given advertisers ongoing desire for relevance and focused targeting.

However, it is important to be aware that not all contextual advertising is created equal. As marketers continue to look for high impact alternatives to their advertising there is an even greater need to understand what ‘contextual advertising’ actually means.

What is contextual targeting? 

Contextual targeting is a method utilised to place your adverts within a relevant and appropriate “environment”. Indeed, a good ‘environment’ provides a good starting point; it will help with brand integrity and reach but a good environment could simply be running the advert on the homepage of a targeted website.  

So, at Alphix Media Exchange, we believe that that definition of environment is too broad, and that contextual targeting should only focus on relevant content (articles, thought leadership, blogs etc.)  

For example, if you were advertising tennis shoes you would want to place your adverts alongside an article about the best Wimbledon Champions of all time. The likelihood of the reader being a tennis player is of course high and there is therefore a higher likelihood they may click on the advert. If we place our tennis shoes adverts next to irrelevant articles on quantum mechanics, the likelihood of a click will decrease. 

Within the contextual space, some solutions target homepages or sections, but not everything within a homepage or a site section may be relevant to your advertising. Depending on the objective of your campaign, contextual targeting might be a better fit than homepage or section targeting. 

Homepage targeting: Running only on homepages of sites – excellent for brand awareness and the feel of a larger campaign

Section targeting: Targeting specific sections within the site (e.g. finance, sport, lifestyle) – excellent for alignment to product themes

Contextual targeting: Placing ads against content or articles that are relevant to the product or topic that is being advertised

Based on research, we estimate that 20% to 30% of ad impressions served via contextual providers are simply served on the homepage of a site, or just a section of a site. 

Our approach: contextual meets deep semantic AI 

At Alphix Media Exchange we believe in precision-based contextual, only identifying the most relevant articles to your advert. We do that by understanding true meaning, not just matching keywords.  

Semantix, our contextual targeting tool, is powered by the best of lexical search, with intelligent semantic search. We use an AI-led approach, utilising both vector-based and lexical-based search algorithms in order to semantically match our search phrases (not keywords) across all articles online. 

In technical terms, semantic search involves converting text into vectors – sequences of numbers – that encode everything about that sentence, from named entities to meanings and even sentiments. This is done using neural networks. Generally speaking, the larger the neural network, the better the embedding. Semantix is powered by embeddings from Microsoft's largest open source, multilingual model. 

In simpler terms, our tool has the ability to fully understand the content itself, identify synonyms, context and semantically match it against our search phrases. 

As we scan across millions of articles, it allows us to advertise against only the most appropriate pages. 

How contextual advertising works in practice  

Let’s look at contextual advertising in practice through an example of an asset manager looking to promote their new bond fund. The aim is to run against content related to bond investing, fixed income, general investing and so on in order to drive traffic to the landing page and align the advertiser with bond investing in general.  

Standard keyword targeting may include terms such as ‘bond’, ‘bonds’, ‘money’, ‘investing’ as well as others. However, all of these terms have more than one meaning. 

• Bond/bonds – this could relate to covalent bonds, or James Bond, bonds of friendship, bonding of objects 

• Money – this could be related to ‘how much money did Taylor Swift make in 2024?’ 

• Investing – this could be related to ‘investing in your wellbeing for greater mental health’ 

This shows the potential for vagueness and irrelevant matching. Much of the targeting will be correct, but when budgets are tight, there is no room for wastage.  

We therefore utilise phrases, for example: 

• “How does market volatility impact bond investing?” 

• “What are the best bond funds for investing in 2025?” 

• “Is investing in bond funds good for my long-term investment needs?” 

This deeper phrase-based analysis immediately gives us a higher chance of placing ads against the most relevant content. Combined with having an in-depth AI-led understanding of all the pages within our chosen sites, we can limit wastage entirely.  

Other mechanisms in market mostly utilise keyword-based algorithms that are ‘lexical’-based. This means they are focused more on just matching words within an article as opposed to a wider understanding of the text/content. This potential for a lack of granularity prompted us to choose an AI-driven approach.  

Why does this matter for you? 

As an advertiser it has never been more important to know what you are actually buying when you buy contextual. You deserve transparency, reassurance and confidence that your message is reaching the right person, at the right time next to the right article.  

By utilising advanced technology, you get more value through better placements and better audiences, likely leading to more engaged readers and better outcomes.  

So when you are looking for a contextual targeting solution, challenge providers, ask how context is defined, find out where ads are placed and what intelligence is behind it. 

The future of contextual is going to be more and more intelligent and sophisticated. You should expect more from your partners, not only at the targeting level but also at the data, reporting and output level. At Alphix Media Exchange, this is what we strive for. Sophistication not only at the front end but also the back.  

By providing deeper level contextual targeting, we run ads for you that provide data points that tell a story. If you want to hear more, please get in touch.