Contextual Buying in Programmatic
With audience based targeting becoming more difficult in programmatic due to a mixture of data legislation such as GDPR & the 3rd party cookie slowly going away on web browsers, contextual targeting has been labelled the futureproof way of running targeted programmatic campaigns. But what does that actually look like today?
Contextual wise you have 5 core options in programmatic:
Google via Google Ads / DV360 which is their proprietary keyword targeting
Oracle (Grapeshot) which is integrated with most DSPs: definitely most popular
DoubleVerify which is integrated with some DSPs
Peer39 which is integrated with some DSPs but has lost its way with Sizmek selling it off
AdmantX which was acquired by IAS in November
Within these, there are different capabilities. All have the ability to target or exclude on a keyword basis but other areas which are limited to certain ones include:
Keyword type (exact match etc) : Grapeshot is king here.
Semantic targeting (positive or negative context) : DV & IAS are strongest here.
Other data points : Oracle utilise Moat for exposure based signals etc
Reporting: Google have ability to report on keyword level, the rest requires bespoke integrations with the DSPs to get it.
Then there is the price. Google is free and always will be. Next is generic off the shelf keyword lists which are anywhere between $0.1 to $0.3. And then custom keyword lists are $0.3+.
In terms of utilising keyword lists in programmatic, the wrong answer is to copy & paste your PPC keywords into programmatic. Whilst some of them do work, the methodology is different and PPC keywords like spelling mistakes do not work the same way in programmatic. This is the best practice each of those vendors would tell you. 20+ keywords works well but the more you have, the better you scale as you would expect.
Contextual targeting is alive & well in programmatic but it is important to understand the capabilities within it when planning or buying on a keyword basis.