Theoretically, any type of action a consumer does online can be defined as their “online behavior.” From their previous actions (i.e., clicking on a banner) to taking those actions and building a predisposition model based on their “likelihood to.” Standard targeting tactics can increase the value of advertising for publishers, networks, agencies, and brands; however, by optimizing the way we target users online, there is the opportunity to increase consumer relevancy and subsequently, conversion rates. It’s the more advanced targeting tactics (those that are beyond standard content and geography) that allow you to reach users further down the purchase funnel.
According to a research study by the Network Advertising Initiative, users who clicked on behaviorally targeted ads increased their likelihood to convert by 2.48 times the conversion rate of standard RON (define) ads while increasing the cost of these ads by only 2.08 times the cost. While the study admits to missing the key component of how likely users are to click in the first place (especially as this pool of online users is continually shrinking), it’s more important than ever to recognize the overall benefit of creating relevancy for consumers.
While standard behavioral targeting methodologies vary across different publishers and networks, it’s not necessarily successful across the board, and results can vary pending overall success metrics.
In the ever-evolving competitive nature of online advertising and opportunity to leverage backend technology, optimizing and trying different types of behavioral targeting are important. Below are a few worth considering and when not to give them a try:
- Standard behavioral: This is based on past behavioral actions inclusive of what ads a user has clicked on, content/pages they have viewed or interacted with, keywords searched, etc. This varies by publishers and networks depending on what they consider and have access to in their behavioral considerations; however, it’s important to find out what provider is being used for the technology and understand the methodology. Historically, it’s worked well in driving quality traffic when the targeting is specific to the brand offering.
- Creative: This is based on a user’s exposure to a specific piece of creative. It can work well when leveraged to expand reach to a specific audience that responds well to a specific targeted site or placements; however, you want to increase frequency. For example, by placing a pixel in a piece of creative targeted on a very niche top-performing site or placement, you can creatively re-target to that user.
- Sequential: This is learning off of consumer’s preferences by their interaction with a piece of creative, and then serving creative targeted to those preferences. It typically works best when the original piece of creative is rich media or has more opportunities for interaction.
- Keyword: This is based on targeting from online users’ previously searched keyword terms. It performs best when used on a category with extremely high or competitive keywords, thereby leveraging a more cost efficient exposure to the same audience through display ads.
- Data driven: This uses backed appended data such as offline spending or HHI and layers this on to any standard targeting. It’s worth testing; however, be warned, as it typically doesn’t scale to high spend or reach levels and can be expensive.
- Retargeting: This is based on a user’s action with the brand website and often performs comparable to brand keywords. This is similar to brand keywords in regards to scalability. Brand traffic directly correlates to how much can be spent each month.
- Lookalike: This is building lookalike models off of the above site retargeting. However, it has not proven any substantial mass success as it’s still in its infancy with many publishers and networks.
While the above list is not inclusive (each company has its own favor), it can be helpful to consider some before excluding advanced targeting tactics from your media buy. As noted in many of the options, often the top-performing metrics are less scalable, hence the importance of testing and optimization.