Contextual targeting may not be a new concept, but it’s one that’s received an increasing amount of attention over the past couple of years. Here, we explain why this is the case and our multi-layered approach to harnessing its power
What is contextual targeting?
Contextual targeting describes a way of delivering advertising that’s primarily based on the context of the online environment in which it’s viewed.
So, imagine you’re looking to plan your next holiday online, and you land on a website with package deals to a particular destination. Contextual advertising allows adverts served on this page to be related to its content.
Alongside deals on flights and hotels you’re viewing, for example, you may see adverts for travel insurance, tour packages, holiday essentials and so on.
How is contextual targeting different from behavioral targeting?
In order to understand the difference between these two types of targeting, it’s helpful to learn the difference between the two common types of cookies: first-party cookies and third-party cookies.
First-party cookies are typically used by a website to remember information about the user in order to improve their experience on that site. This could be their login details, for example, or language preference.
If you’ve ever added something to a shopping basket before navigating away from the page and returning to find these still in the basket, it will have been first-party cookies that ensured this was the case.
They are only used on the website that creates them, hence the name. Third-party cookies, however, are typically created by an ad tech platform, rather than a website you visit, and these are used across multiple sites. This is why you may be used to seeing ads about one website on another that’s completely unrelated to it (a process known as cross-site tracking).
Behavioral targeting makes use of third-party cookies and the above example is a basic way in which this happens. You’ve expressed interest in a product or service, and you will be reminded of this when you’re engaged elsewhere.
Although contextual targeting can make use of third-party cookies, it does not completely rely on them, and does not track users from site to site. Instead, it takes the nature and content of a website currently being visited into account, and combines this with anonymized user information, such as IP geolocation, in order to serve relevant advertising.
Why is contextual advertising receiving more attention today?
The prominence of contextual advertising today is, in large part, a consequence of the default blocking of third-party cookies by popular browser operators.
Mozilla, for example, has blocked third-party cookies from its Firefox browser for almost two years, while Apple introduced the same for its Safari browser through its Intelligent Tracking Prevention (ITP) feature last year. Google’s Chrome browser will follow these in 2022.
GDPR, introduced three years ago, has also played a part. As companies have sought to comply with this – and as browsers have started to block these cookies as standard – the need for different ways of targeting has become apparent.
Quite how the advertising landscape will look once third-party cookies are fully deprecated is not something that anybody knows with any certainty. But contextual targeting, which blends the benefits of a relevant, targeted audience with cookie-free GDPR compliance, is an obvious solution.
As this receives more attention and development – as new technologies step in to automate, simplify and improve existing systems – the need for brands to think carefully about the tools they should be combining to give them an advantage will only become more apparent.
This is necessary as some existing approaches used on their own are too coarse a means of segregating appropriate from inappropriate advertising environments.
One example of this concerns keywords. An advertiser may choose to block adverts from appearing on pages with specific keywords that are deemed to have negative connotations. There’s an obvious logic to this, but going on keywords alone is limiting as it ignores the fact that the sentiment of the page may be completely different (and also the fact that many words considered harmful in a particular context may be perfectly harmless in another).
How does SmartFrame use contextual targeting?
We’ve been developing our in-image advertising technology since 2013, and the result is a more advanced in-image advertising system to the basic ideas described above.
The foundations of this system – understanding site context, keywords and so on – are still relevant here. Understanding the site and the nature of its content, after all, is crucial if targeting is to be effective. But how do we go beyond this?
First, our growing network of images provides advertisers with an ever-increasing range of content against which advertising can be served. As this continues to grow, the network will have greater relevance to a wider range of brands looking to advertise their products and services.
But it’s no use having access to a wealth of images if you don’t understand what they show. And SmartFrame develops its understanding of these on two different levels.
The most important of these is the use of metadata. Thanks to our relationships with photographers, image libraries and rights owners, we can not only access a detailed and diverse set of metadata points for images, but also work with these entities to ensure that future images have the most useful metadata appended.
Much of this may not ordinarily be used by image libraries for anything outside of organizational and archival purposes, but we’re able to leverage this to build a highly accurate idea of what’s going on inside an image. This in turn can be used to better inform targeting, far beyond what ordinarily metadata points would allow.
The other level is the use of AI for scene, subject and sentiment recognition. This is clearly useful in the absence of metadata that one would expect to be manually added to an image.
Image analysis using such tools is obviously useful for detecting any sensitive or problematic content, although current technologies can only reveal so much, and they cannot be expected to identify specific individuals, locations or other elements to a reliably high degree, things that only the photographer or image library would be aware of. So while they can play an important role, their limitations also highlight the importance of being able to understand images through access to comprehensive metadata.
Having an appropriate ad for the right audience is one thing, but for optimum performance, it also needs to be placed where it’s most likely to be viewed.
SmartFrame’s solution combines two key factors to ensure this is the case. The first of these concerns the placement of the ad. Ads are served within SmartFrame images, and these will typically be embedded within the body of an article (rather than someplace else, where banner blindness may be more of an issue). Rather than being served alongside content, these ads are served where the viewer’s attention will already be focused.
Naturally, it’s possible for a page to contain a number of SmartFrames, and serving advertising in each one would be unwise for several reasons. SmartFrame’s approach is to select the SmartFrame in the most prominent position to carry advertising and omit it from any others.
The other key factor concerns viewability. Ads are not only served once the SmartFrame image is in full view, but they occupy the whole of the SmartFrame itself, rather than just a small banner at the bottom. Although the viewer is still free to close these if they wish, it ensures that they will see them as clearly as possible when they first notice them.
Finally, in order to make the advertising more relevant to the audience, location data derived from the user’s IP address can be used to fine-tune targeting. As this data does not identify specific users – rather, a general location within which the user’s device is connected to the internet – it delivers the benefit of more localized targeting without requiring the availability of personal, identifiable data.
Want to find out more about bringing the power of in-image advertising into your site? Get in touch with us today