Digital advertising revolutionised promotional marketing. Instead of the broad-brush ‘one size fits all’ approach of traditional billboards and TV intermissions, advertising in the digital age has enabled much more tailoring to the individual. This can make it more effective, because the audience is being served ads that are more likely to fit their tastes and needs. But there are multiple ways of tailoring, and one of the strongest emerging methods for podcast creators is contextual.
In this feature, we’re going to explain what contextual ads are, how they differ from behavioural ads, and how you can harness their power.
Behavioural versus contextual targeting
Around two thirds of digital advertising uses behavioural targeting. This is the much-maligned system where your activities across all websites, apps and online services are tracked (if you let them), and this data is used to deliver ads. If you’ve spent time in the shoe section of an online store, for example, it’s no longer a surprise that you will see advertisements for shoes elsewhere – and often for the same shoes you purchased a few weeks ago.
Contextual targeting works differently and could be considered more benign than behavioural targeting. Instead of spying on your activities in other locations, contextual targeting uses the topics of the current content to personalise the advertising messages you receive. In other words, the ads are based on what you are consuming at the time, not who the advertiser thinks you are from past actions. On a website, this could be based on keywords within the site content. More advanced forms of contextual targeting can employ AI to understand the meaning of content to match advertising in a deeper semantic fashion, which is where the technology is becoming increasingly valuable.
At this stage, you might be asking how this is different to classic forms of pre-digital advertising, where advertisers match their promotions to print or broadcast media based on its content. However, in the traditional approach a human being will have assessed the publication (or radio programming) and matched this with an advertising message. The difference here is that what is being served can be programmatic advertising, where the ads are placed dynamically in real time by matching the content context automatically with what advertisers wish to advertise against.
One of the reasons for renewed interest in contextual targeting is the increased concern about tracking mechanisms online and their effects on privacy. Web browsers now offer built-in blockers to prevent ads tracking, as do recent versions of Apple’s smartphone iOS. Google also planned to end support for the cookies used in tracking by early 2022, although the phase-out has subsequently been delayed until 2024. With around two-thirds of browser usage being Google Chrome, this would effectively kill cookie-based ads tracking stone dead.
Contextual targeting has fewer privacy concerns, meaning that it is less likely to be affected by new protective measures. It doesn’t need to know who you are to deliver focused promotional messages, because these are tailored to the content instead, and that can still have strong implications for whether you are the right target for an ad. Contextual targeting is also particularly relevant for podcasts, which never had the same levels of behavioural information available as websites and social media platforms in the first place. The nature of podcasts, which are delivered using RSS feeds from a hosting platform, through a directory, to a variety of podcast-playing devices and software, render the typical behavioural signals problematic to obtain. There has never been the same ecosystem of cookies that enable user tracking across devices and services. It’s therefore hard to build the identity graph required by behavioural targeting for a podcast.
Finding the podcasting context
All advertising is aiming to match a promotional message to the right audience. The identity graph is one route to achieving this. But contextual targeting can provide a route to a similar end. Podcasts offer a rich range of content and subject matter, potentially giving advertisers access to finely tuned specialist audiences. Compared to a blog or news website, the context of podcasts was less accessible because audio wasn’t visible semantically. The written series and episode synopses from the RSS feed don’t deliver sufficient detail, and if mere keywords are used, the meaning could be radically misinterpreted. A podcast description involving the word ‘crane’ could be about a bird or mechanical devices for lifting objects.
But the advent of ubiquitous transcription services has opened the detailed content of podcasts as text. This enables the application of Natural Language Processing (NLP) and AI to elicit deeper meaning, which can fuel semantic contextual targeting. Beyond merely the linguistic level, sophisticated AI can interpret the tone and sentiment, for example whether a podcast is serious or humorous. It can also tell whether content is risqué or straight-laced.
This means that advertising technology can now place ads that fit the content of a podcast precisely. For example, a podcast episode that discusses collecting rare trainers could be targeted with advertisements from niche sportswear brands. A company that wants to promote a more lighthearted image can place its messages alongside a podcast with a similar tone. These ads can be delivered programmatically, with Acast for example offering access to topics via keywords and phrases. Acast calls this Conversational Targeting, because it rolls in additional features such as audience intelligence alongside the contextual capabilities.
Deploying contextual targeting on a podcast
The primary way to take advantage of contextual targeting in ads is through your podcast host such as Acast, as described above. This means the ads are added to the podcast via ‘dynamic insertion’ before it is delivered by the host to a directory. In a similar fashion to the ads that appear at the beginning, end and sometimes the middle of YouTube videos, these ads can be programmatic.
Podcasting directory platforms themselves may also offer a similar option. Spotify Ads support contextual advertising, although subscribers won’t hear these, only ‘standard members’ by default. The company launched AI-powered contextual targeting in 2021, enabling episode-level topic-based advertising for brands. Apple Podcasts, of course, doesn’t support ads yet, but some have speculated that it may happen soon and provide a huge source of revenue both for Apple and podcasters alike.
Until then, there are already plenty of opportunities to take advantage of contextual advertising with the right hosting partner. As we pointed out last year, podcasting advertising revenue had a very healthy 2022, and is expected to grow further by 8.1% in 2023. Contextual advertising has a major role in enabling that success. In a market where traditional behavioural methods look increasingly embattled, advertising that can target audiences by context have a bright future, and the focused, engaged podcast listener sits right at the centre.