If I am selling custom-built laptops that are primarily used for gaming, I want to serve compelling ads to gamers at a time when they are likely to be influenced by my messaging.
The most conservative form of contextual display advertising, where we have the greatest control over placements and, therefore, the greatest confidence in the quality of our audience, is referred to as Managed Placement targeting.
Managed placement targeting is a method you can use to specifically choose websites, videos, and apps that are part of the Google Display Network, where you’d like to show your ads.128 Unlike other targeting methods, you select managed placements yourself, opposed to keyword, topic, or audience targeting where your ads are placed on sites automatically for you.
In our gaming laptop example, we can leverage managed placements by collecting a list of websites that include the latest news and information about the gaming laptop market. We can assume that users who are visiting these websites are in the market, or will soon be in the market, for a new gaming laptop.
We would manually select these placements and target them in our Google Display campaigns. This is contextual targeting at its best, but it’s also limited to the quantity of traffic that these sites acquire.
A second contextual targeting method is done through targeting Topics. Topic targeting makes your ads eligible to appear on any pages within the Google Display Network that have content related to your selected topics. As content across the web changes over time, the pages on which your ads appear can now change with it.129
Topic targeting saves you a bit of time, as Google does the legwork for you of aggregating the relevant placements.
Additionally, you can layer your managed placements and topic targeting together, creating even more niche categories. For example, we could choose newyorktimes.com as our placement and Gaming as our Topic. As a result, our ads will only be eligible to appear on NYT articles that are relevant to the gaming topic.
Topics have their own subcategories. One of the subcategories inside the gaming topic is Simulation Games > Business and Tycoon Games, which makes me fairly certain that I could use this option to target articles written about Rollercoaster Tycoon, the game that taught me just about everything I know about business, marketing, and corporate financing.
The broadest contextual targeting option is Keyword Contextual Targeting. In this case, you will choose keywords to match with content throughout the GDN. This is broader than topic targeting but involves more risk. The keyword gaming laptop could match with random placements that have little in common with my target audience, but that is often a risk that we are willing to take—and we can monitor performance over time and make adjustments as needed.
However, as automation becomes more powerful, the chance of keyword targeting showing your ads on unrelated sites becomes less of an issue.
It’s also often the case that the best contextual placements (through managed placements or topic targeting) are extremely competitive and, therefore, expensive. My competitors in the gaming laptop space are likely to also be targeting those exact same placements. After all, this strategy isn’t exactly rocket science—the targeting options are staring you directly in the face.
Similarly, a new fantasy football app is going to have to pay a premium to serve an ad on espn.com, a member of the GDN, because every other fantasy football app and sports-related brand is also fighting for that impression. Inventory, the quantity of available impressions on these placements, is limited, and fierce competition ensures that the cost is well above average.
Therefore, I would argue that advertisers shouldn’t be so selective about their ad placements, especially if they want to scale their campaigns. As a reminder, we need to do two things: validate our audience and get the timing right. Thankfully, we can leverage machine learning to accomplish both tasks.
For contextual campaigns, we assumed that a user was in the market for a gaming laptop because they were browsing one of our selected websites. We used the placement itself to form an assumption about the audience. If our target audience was likely to hang out on those placements, then we would choose them as placements.
Google’s machine learning can predict our target audience at scale and select the ideal time to serve our ads. The only reason we chose those specific placements was because we assumed that the context of the placement itself correctly identified our audience and established that the timing was right. We’ve assumed that if someone is reading reviews on gaming laptops, they might be minutes away from making a purchase decision.
If we can validate our audience by other means, which machine learning can do, we don’t need to be so selective about the actual placement. What’s more, Google can use other signals to predict the proper time to serve an ad, given that we no longer need to make assumptions about when our audience is actively shopping and nearing their purchase decision.
It becomes unnecessary for us to strictly serve ads on gaming websites or other placements that seem relevant to our business. Behavioral targeting allows us to profitably scale our display campaigns to serve ads in placements such as the weather.com app or as YouTube pre-roll before a music video.
Behavioral targeting on the display network begins with Audience Targeting, which was covered in Chapter 5. Available audiences include users you’ve sought out based on your own data (Remarketing, Similar Audiences, etc.) and Google-managed audiences (In-Market and Affinity Audiences).
Keyword targeting on the display network previously had two options. The first was contextual, as described above, and the second was behavioral. Advertisers had the option to choose keywords, which Google would match to audiences. This has evolved over time and is now referred to as Custom-Intent Audiences. Custom-Intent includes users who are searching for, or likely will be searching for, your chosen custom-intent keywords. This is essentially a more advanced variation of the Keyword Targeting option for display campaigns.
