Contextual Targeting on the Google Display Network has always been a bit if a tricky business, especially if you use multiple targeting options for a single adgroup. The new display network tab makes it a bit easier to visualize your selected targeting methods, but which bid was used, or which targeting method generated the context in which your ads were shown is still… shall we say, not quite clear.
Apparently, there is a targeting method hierarchy which uses your “most specific bid” to determine the ad auction. However, the documented hierarchy only extends to two targeting methods at a time. If you use keywords and another targeting method, the keywords determine whether or not the context of the page matches the context of your ads. Then if the content of the page also matches a topic, interest category, placement, or remarketing list, the bid for that secondary targeting method is used. What is not clear, is which bid is used if all targeting methods are part of the same adgroup?
So how do you structure your campaigns and adgroups to ensure that the most relevant, or profitable targeting method is the targeting method used to generate the context in which to show an effective ad?
Perhaps to get familiar with the new way of thinking regarding the display network, would be to only use one targeting method per adgroup. Viewing the campaign from the adgroup level, you can quickly see which targeting methods overall provide the greatest return.
If keyword contextual targeting is not producing a profitable ROI, then DO NOT add keywords to ANY of your profitable display adgroups. Keywords are apparently the top of the pyramid, primary contextual targeting method for any adgroup in which keywords are added. If keyword contextual targeting isn’t working for you, avoid keyword contextual targeting all together!
Let’s say the targeting method that generates the greatest return is topic targeting. If you want to try and optimize to improve your return, first and foremost analyze your automatic placements. Just like you do with keywords for search networks, add profitable placements as managed placements, and/or exclude unprofitable placements. If you add another targeting method (not using keywords) then unless you conduct very controlled experiments over time, you won’t know for sure which targeting method is now the primary targeting method to generate the context in which your ads are shown.
The Google display network still all comes down to the placements on which your ads are shown. Once you establish the targeting method to provide the context, optimize the placements to improve your return. You will find out very early in the campaign level placement analysis a few placements that – regardless of the contextual targeting of the page content – generate the greatest return, or result in the biggest losses.
Regardless of the context of any particular display network approach or goals used, the key to achieving those goals on the Google Display Network lies in the placements on which your ads are shown. That is the only context that really matters!