The modern ability of advertisers to target online audiences, no matter where they are or what they are doing, has prompted many advertisers to disregard the relationship between an ad and its adjacent content in recent years.

Now, however, technological and regulatory shifts are prompting a reappraisal of the tactic. With cookies being retired by 2023, advertisers are seeking targeting alternatives – but solutions from IAB and Google have failed to impress.

With the clock ticking, there’s a growing sense that workarounds approximating the status quo are not quite the answer the industry needs. If the industry is to build back, it needs to be better, and different.

1. Automated, abundant adjacency

Contextual targeting, which seems to tick those boxes, is the practice of matching an ad to relevant content, rather than a relevant user.

Of course, this has long been a feature of traditional print advertising, which has always sold inventory against topical content adjacencies.

In digital, however, semantic analysis algorithms can now do this at scale by examining content like news articles, even videos, to produce keywords describing the topic, emotion and often sentiment contained within. These are fed into programmatic platforms as buyable attributes. It gives performance and scale within a brand-safe model, and can even be used as a proxy for audiences. In short, it’s cookieless and it works.

2. Environmental context boosts viewability and attention

But advertisers should know that “context” can describe more than just the content. Increasingly, it is also being used to describe the whole of the surrounding page.

Viewing advertising on its own is never enough. For it to be effective, users need to actually see an ad – and the context in which viewability takes place is a key influencing factor, as eye tracking research agency Lumen has demonstrated.

In other words, the context of the media environment and its levels of user trust, as much as the relevancy of the content, all work towards influencing attention.

Not all contextual ad-tech is equal

With interest surging in context, and new tools allowing for its scaled use as a means of targeting, marketers can expect a flood of pitches from demand-side ad platforms (DSPs). When sourcing a vendor, however, it will be important ad buyers look under the hood and know which questions to ask to meet their needs.

For example, does the ad-tech on offer only take a keyword approach to contextual targeting – which works, but is a blunt tool – or can it understand sentiment, emotion and semantics, as more sophisticated technologies can?

It is also useful to understand whether the tech has been built from the ground-up, solely for the purpose of delivering contextual advertising, or if it has been retrofitted from something else, such as a brand safety keyword tool. Similarly, is the tech proprietary or off-the-shelf? If it’s the latter, due diligence means investigating not only the software vendor, but also any third-party whose underlying language processing engines may power the product.

Speed test

A good proportion of programmatic contextual targeting works by discovering new content, analysing it, and then activating it for targeting within a DSP. However, as data is gathered and moves throughout the supply chain, it can suffer from latency issues. Reports of 24-hour delays between discovery and activation are not uncommon.

Brands that require real-time customisation in creative, planning or buying are likely to find better performance if the tech is plugged directly into the programmatic supply chain, and that is likely to limit partnerships to more boutique operations.

Don’t take anything at face value

Similarly, it is important to understand the quality and provenance of any data used in the process, as well as the way contextual classifications are made. If an advertiser requires a specific segment, then it really should be specific to them – and not just translated into the closest off-the-shelf template. Advertisers should be clear about their own needs here, and seek a solution that fits them best.

If in doubt, ask a vendor to see examples of how a segment has been defined and how a piece of content has been classified.

Speak with publishers

Contextual advertising works best in higher-quality environments, set against trusted and engaging content on uncluttered webpages. For this reason, it’s also worth building publisher-direct relationships – or going through the joint publisher sales house, the Ozone Project – if there is any concern about contextual ads being served on the long-tail of the web, where we know adspend can fall into a black hole, as this joint ISBA/PwC: 15% of programmatic supply chain costs ‘unattributable’ shows.

Ultimately, context has two sides: relevancy and quality, and both must be considered for it to operate effectively.

Understand the limitations

Contextual targeting is not a direct replacement for targeting with third-party tracking cookies.

Without personal identifiers, issues like managing frequency will be more challenging. This means marketers should approach it with a certain degree of caution, and would be best placed to consider its use as part of a portfolio of post-cookie marketing techniques, rather than a complete solution.

However, ad-tech suppliers and publishers are always looking for ways to improve the tech. In future, data modelling techniques may overcome some of these more immediate limitations. As with all post-cookie alternatives, it’s therefore worthwhile finding the time to test, learn and feed back while they’re in development to ensure they really work for the buy-side.

The climate context

Finally, for the advertisers and agencies using carbon calculators to build greener media plans, it’s worth noting that contextual targeting removes some of the heavy processing required in cookie-matching.

In theory, that could mean it is a less energy-intensive process, and may be a better strategy to adopt if environmental, social and governance (ESG) is a factor.