Years ago Google introduced a product that paired low-funnel high-intent sites with contextual keywords. This GDN product was retired, only to be re-released this year in the form of beta, DSK Campaigns.

What is Google offering in this campaign type?

DSK stands for Display Select Keywords, meaning you only target the most promising keywords and when you take the Content plunge, Google will pair your target with the placements that most frequently lead to post-click conversions.

Like we said, the placement relevance is the promise. High-quality sites for only the most conversion-minded of customers. This means we’re targeting a smaller scope of content, especially when we add in the contextual keywords we deem to be most fitting. You may see a dramatically lower volume of placements, simply because Google’s holding out for just the right fit for your traffic.

The two parts of the DSK campaign are keywords and content targeting. Again, the placements are on Google’s list to take care of while the keywords are up to you. It’s advised that when setting up this sort of campaign, you focus on your top performing keywords and then add in this oh-so-valuable feature:


Opting into “Content” as your keyword setting is the key to enabling this feature. The rest is magic by Google.

The appeal of DSK campaigns is that it’s a low-risk step into the Google Display Network. If you have limited budget to use, this is a good option (because of the aforementioned limited volume). If you are nervous about throwing money into the mystery black hole that can be the GDN, this is also a good option (because of the aforementioned lower spend needed here).

I’m not afraid to give new things a try and the appeal of these campaigns was too great to resist. So what did I do and what did I actually see happen?

First I reviewed the keywords that brought my account the most success. I narrowed them into a relevant ad group and built a campaign that targeted just one region. This was my way of keeping the low risk really low and ensuring I didn’t mess with any of my major regions.

*This is a reminder that everyone takes precautions when trying new things. Be intentional with your spend. Be intentional with your targeting and what audience you’re subjecting to your test. Minimize the risk but maximize the insights.*

My test compares to similar campaign types:


My question was : What kind of limited reach will I see? Will the results be that much better? Will the placements actually be with content I don’t already access? Should I keep running this or, better yet, apply it to more regions?

The results I have seen so far are telling me the right things, albeit interesting things.

Spend: I started this campaign with an extremely limited budget. After a day or two, the engagement (read: CTR) was amazing. SO amazing, I increased my budget to tempt Google (and yes, fate) a bit more. After increasing my daily spend by 5x, I still see that familiar “Limited by budget” message coming through. Dangit, Google, I’ll bite. I’ll go ahead and push my budget just a slight bit more. For now.

CPCs: After just a few weeks, the average CPC for the DSK campaign was 67% lower than that of the traditional Contextual + Topics. So the volume I’m able to bring through this more relevant is more cost efficient. More for less? Yes, please.

Audience Reach: Of those customers I’m now showing to, what’s the current engagement like? The current CTR for the DSK campaign is 60% higher than the Control campaign, and the average session duration for these visitors was indeed 1.5 times greater from the DSK campaign. The percentage of New Users coming through the DSK content is also about 8% lower than the Control campaign. This means that those who are being tied to the Display Select targeting may have in fact visited our website before, a nod to ensuring a strong remarketing strategy.

Placement quality: The DSK campaign indeed amassed 95% fewer placements than the Control campaign, a result we expected to see.


Interestingly enough, 15% of the DSK placements were also present in our Control campaign. So perhaps we were reaching these folks on these sites already and just didn’t know it. Of the DSK placements we’ve seen, those that overlap with the Control placements actually have a lower engagement than the other DSK placements.

Final aside on placements: Because the DSK is a fresh start, we were less restrictive of mobile activity. From this we saw an even higher average CTR from mobile app placements. These apps were extremely relevant to our customer base and will remain within our placement eligibility – an outcome not all GDN campaigns see.

Going forward, I’m going to continue to push this campaign. I may end up backing off the Control campaign if cost efficiency, engagement and onsite behavior continues to be strong. I will continue to watch how my placements perform and where I see overlap, should I ultimately exclude a particular placement from a specific campaign. More than anything, I do plan to expand my DSK efforts to remaining regions with the expectation that I’ll continue to see relevant reach and strong engagement.

At least, until Google takes this beta away too….