In an increasingly privacy-conscious web, contextual ad targeting is having a moment.
Discussion website and self-proclaimed ‘front page of the internet’, Reddit, recently announced the acquisition of Spiketrap, an audience contextualisation company. As well as further building out Reddit’s advertising offering, the move to acquire Spiketrap highlights Reddit’s intention to capitalise on advertising that matches a user’s interests – a powerful proposition for a platform with more than 100,000 active interest-based communities.
Likewise, TikTok introduced a new contextual advertising solution in May, TikTok Pulse, that enables brands to appear next to top-performing content on the all-important ‘For You’ tab. Google’s quest to find an effective replacement for intrusive third-party cookie tracking and targeting has also led it to focus on something known as ‘Topics’, which assigns users a set of interests based on the websites they have visited, a move that has prompted comparisons to contextual targeting. However, as Erin McCallion, Chief Marketing Officer at contextual intelligence company GumGum points out,
“Google’s announcement of Topics has raised interest in contextual advertising, but it’s important to remember that Google Topics isn’t contextual-first, because it ultimately relies on behavioural data. With data privacy regulations constantly evolving, Topics could well face regulatory hurdles in the future.”
So, what exactly do we mean when we talk about contextual advertising? Why is this particular type of advertising – once the predominant method of online ad targeting, which later took a back seat to behavioural targeting – experiencing a resurgence in popularity? And how can brands employ it to best effect?
This briefing covers:
- What is contextual advertising, and why is it seeing a revival?
- How does contextual advertising work in 2022?
- What are some advantages of contextual advertising?
- Data quality is still key
What is contextual advertising, and why is it seeing a revival?
Anyone who was an internet user in the 1990s or early 2000s will likely remember the way the advertising that appeared on a webpage would typically match the content of the page around it – or, to put it another way, the context it appeared in. A user browsing a car enthusiasts’ discussion forum could expect to see advertising related to cars, or car parts – the assumption from advertisers being that anyone who visited such a discussion forum would logically have an interest in buying cars, or car parts.
This model for digital advertising was nothing new – rather, it was simply the online equivalent of a model that print media had been using all along. A diving magazine, for example, might feature an ad for diving holidays on its inside front cover – or partner with a diving accessories provider for a deal.
Indeed, a notorious early example of online advertising – sometimes credited as the very first banner ad – was an ad placed by telecommunications company AT&T on HotWired.com (the digital counterpart to Wired magazine) in 1994 that touted the benefits and innovations of the emerging internet. Again, it could be reasonably assumed that someone browsing the digital version of a technology magazine in 1994 was part of AT&T’s target demographic.
However, the emergence of third-party tracking cookies led to the birth and growth of an advertising ecosystem founded on tracking users’ behaviour around the web – rather than limiting itself to the site that they were visiting in the moment – which was a very attractive prospect to advertisers. So much so that, in the decades since that early banner ad, it has become difficult for many marketers and advertisers to conceive of an industry without third-party cookies.
At the same time, a growing movement towards protecting users’ privacy online has also led to a backlash against third-party cookies and the advertisers that make use of them. Consumers are increasingly aware of, and averse to, the large amounts of data collected about their habits and used to fuel advertising: research by independent adtech company MediaMath published in January 2022 found that 51% of US consumers were “not comfortable with websites tracking online behaviour and capturing personal information such as the other websites [they] visited.”
Web browsers have responded by implementing third-party cookie blocking by default: in 2017, Apple’s Safari introduced Intelligent Tracking Prevention (ITP), which prevented cookies from being used in a third-party context after 24 hours if the user hadn’t visited the originating site, while in 2019 Firefox released Enhanced Tracking Prevention (ETP), which blocks third-party tracking cookies and cryptominers, later enabling ETP by default for all Firefox users. Both browsers have since built on these protections, doubling down on their commitment to allowing users to opt out of third-party cookie tracking.
Chrome, the world’s most popular web browser, was a much more reluctant follower of this trend (given that it is owned by advertising titan Google), but eventually announced in 2020 that it would withdraw support for third-party tracking cookies by 2022. This deadline was later pushed back to 2023, and then to late 2024, as Google grappled with the complexity of maintaining its advertising empire without third-party tracking cookies. However, for advertisers, this show of support from the largest web browser by market share was a sign that adapting to a cookie-less world had become imperative.
The resurgence of contextual advertising
In the past few years, contextual advertising has experienced a resurgence in interest as marketers and advertisers realise they need to adopt new methods of targeting that don’t rely on third-party cookies. Contextual advertising is one such method, as it instead relies on parsing the content of a particular webpage or piece of media in order to serve the most relevant ad in that context.
“In simple terms, contextual targeting is a method of targeting that leverages the content someone is consuming to target them,” says Ed Blakeway, Head of Programmatic at Journey Further. “It scans the content for relevant keywords which indicate the context in which the ad will be served. Advertisers will pick multiple keywords that people read to proactively target users rather than using cookies and tracking pixels to determine the behaviour of a user.”
Contextual advertising can also be served within media like videos, audio, and images, drawing on data such as sound, speech, image analysis and metadata for context. Contextual ad targeting has proven particularly fruitful in the podcast industry, where the common methods of consumption – through an RSS feed or by downloading individual episodes to a podcast app – make it difficult to employ traditional tracking methods like third-party cookies or mobile ad identifiers.
Contextual advertising within OTT (over-the-top) streaming content may also be coming into new prominence as streaming services like Netflix and Disney+ launch cheaper ad-supported subscription tiers. Netflix’s advertising offering, which is rumoured to be launching as early as November, will reportedly allow ad buyers to target viewers watching a specific genre of show, as well as viewers consuming Netflix’s top 10 shows in the United States. Not to be left out, YouTube – already a major provider of contextual video advertising – also announced the launch of free, ad-supported TV show streaming in March.
Juliet McCutcheon, UK Sales Lead at technology and data platform Channel Factory, which offers contextual advertising on YouTube, said that it is too early to tell yet what the introduction of these ad-supported streaming tiers means for contextual advertising. “It is dependent on the contextual capabilities of the various streaming services and their data collection,” she says. “Only once this is fully understood by the streaming services themselves with their ad-supported platforms, will they be able to demonstrate to brands and agencies the level of context they will be able to offer.” She believes that YouTube will remain a “key player” in this space due to the “healthy level of first-party data” that Google can offer.
McCutcheon confirms that Channel Factory has seen a rise in client interest in contextual advertising as third-party cookies have increasingly come under threat. “Whilst the removal of cookies has been delayed a handful of times, we can still be sure that is on its way,” she said. “Brands and advertisers alike must be prepared for that fateful day … Channel Factory has definitely seen a rise in demand for highly contextual campaigns since discussions around the removal of cookies began.”
How does contextual advertising work in 2022?
Advertising technology has come a long way since the early days of online contextual advertising in the 1990s. Advances like machine learning, natural language processing and semantic analysis enable context to be understood in a far more sophisticated manner than was possible a little under three decades ago.
Today, marketers can specify a target topic – such as cars – and a demand-side platform (DSP) will place their ad on webpages that meet those targeting parameters. Sophisticated contextual advertising solutions may also offer opportunities to target niche contexts, such as “classic cars”, “classic 1960s cars”, or “classic American 1960s cars” for even more precise targeting. Alternatively, platforms may use AI to detect variations of the original target term and enable brands to appear in related articles.
Sentiment analysis also helps brands to ensure that their ads are appearing in a brand safe environment: a blog post about how buying classic cars is wasteful and expensive would not be a good context for a classic car vendor to advertise, no matter how relevant the topic. (We’ll touch more on brand safety and suitability further down). Sentiment analysis can help to detect when otherwise relevant-seeming content isn’t appropriate for the brand, avoiding an awkward faux-pas – and advertisers should also be able to verify that their ads are being served on relevant domains before, and during, activation.
GumGum’s Erin McCallion notes that advances in things like natural language processing and computer vision are bringing contextual targeting abilities to the next level, which she dubs “contextual 2.0”. “This is opening up the potential of content-level contextual analysis that will revolutionise digital advertising,” she says.
“Many contextual solutions still focus on basic keyword analysis. We know that keywords alone tell you very little about the content on a web page or digital environment. For example, a keyword solution might identify content with the word ‘shoot’ in it as potentially about violence, despite the fact that it was actually referring to a ‘basketball shoot’ or ‘photo shoot’.
“Natural language processing [can] bring human level understanding to text and audio-based content and avoid these mistakes. At the same time … computer vision [can] help us analyse and understand image and video based content in the same way a human can.
“The way we see it, Contextual 1.0 sees digital environments in black and white, whereas Contextual 2.0 sees digital environments in technicolour.”
Contextual ad buying in 2022
Programmatic auctions have further sped up and automated the process of buying contextual advertising. Journey Further’s Ed Blakeway explains how contextual advertising space is sold programmatically:
“In the same way an advertiser values a user based on their behaviour, they will value users based on the content they read,” he says. “Typically these CPMs are put to an auction in the same way as any other programmatic auction.”
Blakeway notes that the cost of contextually-targeted advertising tends to be higher than behaviourally-targeted advertising due to the specificity of criteria and the volume of auctions entered. However, the value of contextual advertising lies in the ability to target consumers who are demonstrating high interest in a particular product or category at that particular moment. “Often the audience chosen is a very focused group of high-intent users ready to interact – essentially, you pay a premium to eliminate wastage,” says Blakeway.
Some contextual ad providers still favour the direct-to-publisher model, such as Smartframe Technologies, a provider of “image streaming” technology that offers embeddable images to publications with no license fee. The image itself provides the context for an ad, which is temporarily overlaid on an image – for example, an image of a bed might display an ad placement for Ikea. While Smartframe sells some inventory programmatically, the company favours hyper-relevant contextual advertising in order to minimise the intrusiveness of the ad and ensure a positive end user experience.
“We much prefer offering a complete creative experience,” says Smartframe CEO Rob Sewell – in other words, a complimentary combination of brand, image and publication, such as an ad for condiments appearing over the image of a burger on a food blog. “It’s very hard to police programmatic to ensure high relevance – whoever gets the highest bid is going to get the space.
“If you can really highly contextually target to the content that person’s consuming, then the ad should be just as interesting as the content that it’s appearing within,” he adds. “It’s got to be ads that people want to see.”
The rise of social media since the late 2000s has also presented new channels and options for contextual advertising. Blakeway comments that conversational platforms like Twitter and Reddit are “underrated as contextual tools, as the ability to interrupt or join in a conversation based on particular keywords is very powerful.” However, he notes that while the impact of a contextual message in these spaces might be immediate, “it lacks the scale that the wider web has via Google” – which provides traditional contextual ad placements across its ad network – “or other DSPs.”
One set of social platforms that offer little to no contextual advertising are the Meta-owned trio of Facebook, Instagram and WhatsApp, relying instead on the extensive amounts of behavioural data within their walled gardens to target advertising. However, this has also left the likes of Facebook vulnerable to the impact of privacy-oriented measures like Apple’s App Tracking Transparency, which requires iOS users to proactively opt into activity tracking from individual apps, and reportedly cost Facebook a little over $8 billion in lost revenue in Q3 and Q4 2021.
What are some advantages of contextual advertising?
Keeping pace with the user
We’ve already touched on some of the benefits of contextual targeting, including a lack of reliance on third-party cookies and other identifiers like mobile ad identifiers, reduced wastage, and high relevance in the moment. Smartframe’s Rob Sewell notes that contextual advertising can be more suited to the fast pace of people’s daily lives when compared with behavioural targeting, which can often lag behind.
“What I’m looking at on the internet a day ago – certainly a week ago, or a month ago – is going to be completely different to what I’m looking at next month. So for ads to follow me due to my history? My life moves at a fast pace – if I’m doing some home improvements, and I’ve finished my bathroom, and I’ve moved onto looking at interior design for my dining room, if there’s an ad that will give me some ideas and inspiration for my dining room, brilliant – instead of a pair of taps or a sink, when I’ve got no more bathrooms left to do.
“I think that’s where contextual really plays – because it’s relevant to the given moment.”
Contextual advertising can also be more suitable than behavioural targeting for situations in which multiple users share a device, as the previous behaviour exhibited on the device may have little to do with the person currently using it. This is difficult for advertisers to determine behaviourally, but contextual advertising will reach those users through the content they are accessing at that moment, regardless of what it is.
Erin McCallion adds, “The reality is that, as human beings, we naturally move in micro-moments, floating from one topic to another. Digital advertising should be able to move with us and be relevant at that exact moment, without invading our privacy.”
Brand suitability and safety
Another strength of contextual advertising is brand suitability and safety. The content that an ad appears next to can significantly impact consumer perception of that brand, and appearing next to content of a fraught, harmful or sensitive nature can be potentially damaging for brands. This problem has only become more acute in recent years with the rise of fake news and misinformation, not to mention the additional challenge presented by the Covid-19 pandemic.
Contextual advertising alleviates this problem by ensuring that advertising will appear alongside content that is directly relevant to the brand. However, as alluded to earlier, there are also contexts that fit the right criteria but that would nevertheless be inappropriate for advertising. To avoid this, brands can specify “out of context” phrases to prevent their advertising from being placed alongside this type of content. For example, an alcoholic beverages brand may want to target web content and media related to alcohol (such as cocktail recipes or listicles of the best brands), but would definitely want to avoid placing ads on content about alcohol abuse or addiction.
For advertising on YouTube, Channel Factory offers both inclusion and block lists that can be customised to brands in order to ensure they appear next to content aligned with their values and not dangerous or irrelevant content, or misinformation. “We ingest data from YouTube via API into our platform, ViewIQ, to ensure our clients’ ads run on brand-suitable and contextually relevant content,” says Juliet McCutcheon. “We offer standard IAB category lists” – based on the Interactive Advertising Bureau’s taxonomy for classifying web content – “as well as highly customised, bespoke lists across 49+ languages to ensure our clients … reach meaningful, engaged audiences.”
Sometimes human oversight is required to guard against potential brand safety issues. Rob Sewell explains that Smartframe relies on human-submitted metadata for images rather than computer vision, because computers can still get things badly wrong – he gives the example of a terrorist wielding a machine gun being misinterpreted by Google’s image recognition as a “man with musical instrument”. “The metadata comes from the photographer who’s looking down the lens, taking the picture – he knows exactly where he is, exactly what the subject matter is, and that’s all tagged up as metadata.”
This can have additional advantages for brand relevance – for example, metadata can identify a specific car brand, allowing ads relevant to that brand to appear over the image.
Less is more
Ed Blakeway notes that behavioural targeting is often based on extrapolated data, “so audience size is often disproportionate to the budget. However, this is where contextual targeting excels. Whilst behavioural targeting can be vague, contextual advertising is scalable with minimal wastage. It allows you to hit multiple pockets of audiences and subsections of those audiences based on the content they read.”
Rob Sewell agrees that with contextual advertising, less is often more: “If you can serve a really highly premium ad in the right contextual environment, maybe you need to only serve 10,000 to get the same response [as a larger number of non-contextual ads] – in fact, a better response, and a better brand experience.
“You might have to pay a higher CPM, but you’ll have … a better brand response, and a better connection with the clients you’re trying to engage with. Quality over quantity is the way I believe the future of advertising should go.”
Blakeway adds, “Contextual targeting is future-proof and a sure-fire way of knowing you’re hitting the right audience and the nature of it means you can not only define who the audience is, but influence their decision at each stage.”
Data quality is still key
With targeting methods like behavioural and demographic targeting, quality data makes a huge difference to the effectiveness of an ad – and the same is just as true for contextual advertising. “Whilst contextual targeting can fit into any marketing brief, it has a limit and it’s highly dependent on good content,” observes Blakeway.
Contextual data can also be combined with other types of data, such as high-quality first- or second-party behavioural data, to create an even more relevant in-the-moment message, as Blakeway explains:
“I’ve seen behavioural targeting work well with contextual targeting in scenarios where you can source data from publishers relevant to your content. Accessing traveller demographic data from rail companies, active users on parent forums and delivering contextual ads targeting “things to do with the kids” is an example of how you can target users when they’re actively engaging in relevant content but also whilst they’re browsing.
“The onus is (and has always been) on the quality of data,” he adds. “The previous example outlines how the behavioural users came from a reputable source rather than an amalgamation of random signals online, which could be out of date.”
No type of targeting, contextual targeting included, is a silver bullet, and data quality is always important to produce good results. However, contextual targeting opens up new possibilities for reaching users when they are likely to be receptive to particular types of product messaging, and can be used in conjunction with carefully-sourced behavioural or demographic data to enhance its effectiveness.
“Broader behavioural targeting can also fuel your contextual efforts,” says Blakeway. “Post-campaign analysis from these broader audiences will outline placements, content and categories that your converting customers engage with most. This balance of scale and precision is how you can constantly improve your campaigns and make your ads more relevant.”
Unlocking new possibilities for targeting
Contextual targeting, while applicable across all sectors, can also offer specific advantages for particular sectors. For example, in the pharmaceutical sector where companies might be selling products that relate to sensitive health or personal issues, it is unethical – and may even be prohibited by regulation – to identify and track users who fall into those categories. However, by implementing contextual advertising, companies can still reach users who might be in the market for these products without needing to know anything about them.
Consumer-packaged goods (CPG) brands tend to target a broad range of consumers and have a short sales cycle. Contextual targeting can broaden this targeting still further and unlock new potential audiences – for example, a cosmetics company that might ordinarily target teen girls can instead broaden its targeting to “teen fashion” of all kinds, thereby reaching their parents, or teenagers of all genders. A coffee company that might ordinarily have been behaviourally targeting users who had interacted with coffee-related content could broaden its reach by placing contextual advertising on content for students, complete with a student-specific offer to convert new customers. A company producing pasta could broaden its reach by advertising on content related to Italian culture more generally.
Contextual advertising also comes into its own for major purchases that the user might not have previously demonstrated a history of engaging with, such as a new car – which is likely to be needed rarely – or a new insurance plan. By advertising contextually on content related to new car purchases or car brand comparisons, an auto brand – or car insurance provider – could reach a consumer who is planning a new car purchase right as they begin their research.
While the phasing-out of the third-party tracking cookie ignited a flurry of alarm among many marketers concerned about their ability to continue reaching the right audiences, methods like contextual targeting show that privacy-oriented advertising can be just as effective – and unlock entirely new possibilities for targeting that may deliver even better results. Speaking to Econsultancy in June, Rob Sewell cited contextual advertising as one of the major trends shaping advertising in 2022, noting,
“The necessity for [contextual targeting] is now well understood in the industry, so I imagine the conversation will shift to the specific ways in which this can be harnessed to ensure advertising is relevant and impactful, and for brands to get the maximum ROI on their campaigns. … I think this focus on user experience and gaining maximum attention will be a cornerstone in shaping the future of digital advertising.”
Ed Blakeway adds, “The focus now should be on how we can develop the sentiment of the content being analysed, as well as helping advertisers discover new content and keywords to target based on the content and how successful that has been in driving the desired actions.”
Digital Transformation Monthly: The Rise of Retail Media Networks
Comments