Advertising is about reaching the right person at the right place at the right time. Contextual targeting is commonly used to find the “right place.”
Yet, earlier this year, this fundamental feature of programmatic advertising came under fire when the UK’s Association of Online Publishers (AOP) denounced it and identified it as a catalyst for intellectual property theft. This is a grossly misleading argument.
Contextual targeting works to the benefit of publishers. Of course, the misappropriation of a publisher’s audience data would be unethical. But this is not what contextual providers do.
While publishers might be seeing their open-market CPMs down, they are not alone – and contextual targeting is not to blame.
To understand how contextual targeting differs from theft of publishers’ first-party data, we need to examine how it really works.
What is contextual targeting, really?
Contextual targeting and brand safety targeting are standard features of programmatic advertising, and each is driven by contextual classification.
With contextual targeting, DSPs let buyers bid in real time on inventory that has been classified as contextually relevant to their campaign. When a DSP receives a bid request, it queries the contextual provider with the URL, which then returns their classifications – all for a small CPM fee.
Brand safety targeting, meanwhile, protects buyers from bidding on inventory that is paired with inappropriate, low-quality or irrelevant content. This works to the advantage of publishers by steering campaign bids away from less reputable sites and lower quality content, and instead directing buyers to bid on their inventory.
Contextual targeting is not synonymous with stealing a publisher’s audience and targeting them elsewhere without compensating the publisher. Whether they segment the users into audiences manually, or use contextual classifications to do so, the issue remains the creation of audience data. But contextual targeting doesn’t involve the creation of audiences. It’s simply a real-time decision about whether or not to buy a given impression from a publisher based on content on that publisher’s page.
As such, it’s misleading to assert that the process of contextual classification is improper after it has been freely and openly discussed for more than a decade.
In its accusation, the AOP uses the phrase “contextual audiences” and refers to “packing unseen extra tags” and “first-party data extraction.” This language is confusing, because “audience” usually means a list of users (cookies, mobile advertising IDs, etc.). “Packing unseen extra tags” is a way of extracting these users from a publisher’s site. It does not conceptually apply to performing a contextual classification for ad targeting.
Publishers still have control over contextual classification
As the advertising industry stares down the barrel of a “cookieless” future, contextual targeting is becoming more important than ever. More than half of the campaigns created on my company’s DSP use contextual targeting in some way.
Publishers already have the tools to control contextual classification. In fact, it is only possible for a DSP to offer contextual targeting because bid requests have a field for the page URL. This means that it’s entirely up to the publisher to decide whether this information is provided or not.
Many exchanges and supply-side platforms even have an option to suppress it, giving publishers greater control over their data. Depending on the implementation, SSPs may rely on the publisher to explicitly pass the full-page URL – meaning the publisher could simply choose not to share it if they don’t want any buying decisions to be made based on a classification of their site.
Few use cases exist for a DSP to have a publisher’s full-page URL except for contextual targeting. The publisher’s domain is provided in a separate field, giving publishers the option to share their domain while suppressing full page URLs. In that case, a contextual classification can only exist at the overall domain level, if provided at all.
But while publishers can choose to exclude contextual classification, they could exclude a substantial amount of demand, and therefore revenue, with it. Advertisers want to buy impressions on contextually relevant content from reputable, premium publishers, and they want to avoid inappropriate content.
The path forward for publishers, DSPs and advertisers
As advertising budgets are cut and data privacy protections sweep through governments around the world, advertisers have fewer targeting options on the user level. This makes contextual targeting incredibly valuable.
For reputable publishers, contextual targeting is not the enemy. If a publisher is creating high-quality content that resonates with an advertiser’s target audience, it is more likely that this advertiser will buy inventory on the publisher’s website through a DSP. This means more valuable impressions and clicks for buyers and higher ad revenues for publishers.
While publishers can attempt to force licensing agreements with DSPs to tighten their grip on inventory, this approach would likely end in failure. Negotiating agreements with DSPs would exhaust both publisher and DSP resources and further restrict CPM revenues.
Contextual targeting is a mutually beneficial ecosystem where buyers, DSPs and publishers can work in harmony to gather and deliver the necessary insights and continue to grow our revenues.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow Basis Technologies and AdExchanger on LinkedIn.
For more articles featuring Ian Trider, click here.
