With tightening data regulations and increasingly common browser cookie restrictions, some in the industry predict a return to contextual targeting, at least in the short term. If advertisers find audience data cut off, they may see context as the best alternative way to ensure ads are served to relevant audiences.
But this doesn’t necessarily mean a return to the sorts of contextual targeting we saw in the earlier days of internet advertising – serving coffee ads on coffee related content, for example. Instead we’ve seen a new breed of contextual targeting solution emerge, keen to highlight the sophistication of modern contextual targeting. Here are some of the most interesting companies VAN has seen develop in this space:
Founders: Nadav Shmuel and Christian Dankl
One of a number of contextual advertising companies which focuses on YouTube advertising, co-founder Christian Dankl says Precise TV is different through how it targets individual bits of content, rather than using a top-down channel based approach. Precise TV says it analyses contextual data for millions of videos, and uses machine learning to enable more granular targeting – for example, finding videos likely to resonate with people interested in expensive, artisanal coffee, rather than just targeting ‘coffee’ as a broad category.
Dankl says Precise.TV’s machine learning algorithms also allow advertisers to discover contextually relevant content they might not be aware of. For example, he says that amateur high performance cyclists tend to also have an interest in artisan coffee making, making content relating to high performance cycling contextually relevant for artisanal coffee brands. Precise says its ‘Contextual Data Management Platform’ (DMP) helps identify these correlations.
Founder: Peter Mason
London-based Illuma says it looks at how consumers are interacting with content, and uses machine learning to identify which content is performing well for which segments, in order to classify context. The company says this means it can identify in real time which sorts of content a given brand should be targeting, given their audience.
Illuma says for example that for a yoga brand, yoga related content might perform best over the campaign’s overall time frame. But there will be times where other content categories like health, news, or gardening may perform better. The company says its “adaptive contextual technology” identifies these trends, enabling more efficient contextual targeting.
Founders: Rich Raddon, Zach James
Another YouTube specialist, Zefr says it processes the meta data for every video uploaded to YouTube to determine the context of the video and gauge if it’s relevant for a given brand. The company says it forecasts the number of views and impressions a brand will get from their targeting parameters, and also uses a mix of technology and human review to filter for brand safety.
Zefr says one of its differentiators is its “human in the loop” machine learning approach. The company says it first captures attributes important to brands, including their content preferences, their brand values, and their campaign goals in order to create a custom model to identify which content is relevant for the brand. Doing so requires strong training data, which is driven by humans. This means that humans review videos picked out as contextually relevant and brand suitable by Zefr’s algorithm, to ensure it’s on track.
Founders: Juha Korhonen, Rami Alanko, Tatu Salminen
Beemray sells itself as targeting “human context”, which the company says takes into account the user’s mindset and external factors, alongside the content they’re consuming. Beemray ingests information provided in the bid request including IP address, a lat/long, and the URL to make judgements about the individual’s current situation, for example whether they’re at work, at home, or travelling. It then adds in other external data which might be relevant, like weather, traffic, or sports results, to identify contextual moments which will be most relevant for a specific brand.
The company says brands can also use the data they get to build profiles of the types of audiences they should be targeting, without the need to profile individuals. For example, a particular brand might find their ads work best when served to commuters, when it has just started raining, and their football team has just scored.
Grapeshot (Video Context)
Founders: John Snyder, Martin Porter
Grapeshot is one of the older and better known names in contextual advertising, and its acquisition by Oracle in 2018 is seen as somewhat emblematic of the renewed appetite for contextual targeting. But Grapeshot’s contextual targeting solution for video advertising wasn’t released until 2018, shortly after the Oracle acquisition.
The solution, ‘Video Context’, turns the audio content of a video into text, and analyses that text in order to determine the video’s context. This information can then be used to determine contextual relevance for brands buying in-stream video ads.
Grapeshot say analysing the audio track has advantages over meta tags which are often set by humans, and not always reliable. But the drawback obviously comes in cases where video have little or no spoken content, or the audio doesn’t represent the video’s visual feed.
Founders: Alex Modon, Hitesh Chawla, Mudit Seth
While Indian ad tech company SilverPush has been operating since 2012, its contextual video advertising product Mirrors was launched late last year. SilverPush says Mirrors uses video content recognition technology to recognise faces, logos, objects and emotions within video content. It then uses this information to determine contextual relevance for specific brands, serving in-video ads where it spots of contextually relevant opportunity.
SilverPush says Mirrors could, for example, recognise when a user is watching a video of Cristiano Ronaldo scoring a goal, and then serve an ad featuring Ronaldo endorsing a brand. Or it could spot when a brand pops up in a piece of content, and serve an in-video ad for that brand.
Founders: Field Garthwaite, Richie Hyden, Robert Bardunias
IRIS.TV is another established ad tech company to recently release a contextual ad targeting solution for video, launching its solution at Cannes earlier this year.
IRIS.TV already operated video intelligence tech, which broadcasters and publishers could use to serve more relevant content to their audiences. And the company has added to this the ability to classify editorial content into contextual segments. IRIS TV’s personalisation engine, Adaptive Stream, then passes these segments into the client’s ad server, which enables both direct and private marketplace transactions for pre-roll and mid-roll ads to occur, based off these segments.