Behavioral Targeting: Why Every Ad Server Should Implement It | Adzerk

Behavioral Targeting: Why Every Ad Server Should Implement It

Jane O'Hara
cardimage

Behavioral targeting, also known as audience or interest targeting, is an ad targeting feature that tailors messages to users’ previous actions.

As you build your own ad server, behavioral targeting is a targeting feature you’ll want to consider.

This article explains what behavioral targeting is, how it works, why it’s beneficial, and why you should incorporate behavioral targeting into your ad server.

What is behavioral targeting, and how does it work?

Behavioral targeting is where you retarget people based on their previous interactions with your site or app.

Unlike demographic targeting, which makes assumptions based on a user’s age, gender, and the like, behavioral targeting leverages users’ past actions to indicate their interests.

A travel site like FlyWithUs, for instance, can use their first-party data on past behaviors - such as what searches the user did on FlyWithUs's homepage - to create custom audience segments like “Hawaiian vacation planners”, “budget travelers”, or “spa hotel lovers”.

Advertisers could then pay to target just that segment when the user is back on FlyWithUs's site.

FlyWithUs wouldn’t share PII directly with the advertisers, but they would be building audience segments that advertisers could target at premium rates, such as this illustrative rate card:

Segment CPM (cost per thousand impressions)
Site-wide (no targeting) $1
Frequent Flyers (3+ flights/yr) $5
Business Travelers (10+ flights/yr) $7
Hawaiian Vacation Planners $25
1. Site-wide (no targeting) - $1 CPM (cost per thousand impressions)

2. Frequent Flyers (3+ flights/yr) - $5

3. Business Travelers (10+ flights/yr) - $7

4. Hawaiian Vacation Planners - $25

Here, a Honolulu-based rental car company may be happy to pay $25 CPMs to reach that audience on FlyWithUs's site, since the return-on-investment for that hyper-targeted segment may be high (though volume likely low).

By creating these audiences using your proprietary data, you could greatly increase your ad rates.

What are the benefits of behavioral targeting?

There are three key benefits to implementing behavioral targeting into your ad server:

  1. Adding this feature could generate more revenue, as advertisers will pay more to reach their target audiences
  2. It offers a better user experience, as ads are based on known interests
  3. It gives you feature parity with other ad platforms

For these reasons, we believe behavioral targeting is a must-have feature for your ad server, as it offers multiple benefits: an opportunity for increased revenue, along with better experiences for your users and advertisers.

Do other ad platforms offer behavioral targeting?

They do, so advertisers may expect the ability to target consumer behavior on your platform.

Platform Have?
Google Ads Yes
Pinterest Yes
Snapchat Yes
Facebook Yes
Amazon DSP Yes
DSPs & Ad Exchanges Yes
LinkedIn Yes
Twitter Yes
Quora Yes
Microsoft Ads Yes - limited to retargeting
Reddit Yes
Amazon Sponsored Products No
eBay Promoted Listings No
Etsy Promoted Listings No
Facebook allows advertisers to target many pre-defined segments, such as "Engages In Moderate Political Content" and "Commuters".
behavioral targeting options for facebook
behavioral targeting options for facebook
Twitter also offers behavioral targeting, including the ability to target people who like specific types of alcohol (beer, spirits, and/or wine).
behavioral targeting options for twitter

Another example is Pinterest, which has many interests to target:
behavioral targeting options for pinterest
behavioral targeting options for pinterest

How do I implement behavioral targeting?

Here’s a breakdown for getting there:
Decide what audiences you want to build
Think about what type of segments you want to sell against. This will be particular to your product, but here's a quick illustration of how an events listing website, such as EventsForMe, may approach it:
Segment Description
Broadway Lovers User has made a search for a Broadway show, or clicked on a Broadway listing, in the past year
Broadway Purchasers Has bought a Broadway ticket in the past year
Current Broadway Attendees Attending a Broadway show that week

1. Broadway Lovers - User has made a search for a Broadway show, or clicked on a Broadway listing, in the past year

2. Broadway Purchasers - Has bought a Broadway ticket in the past year

3. Current Broadway Attendees - Attending a Broadway show that week

An advertiser could then choose to target one of these segments, and the ad would appear no matter where the user is on EventsForMe's site.

For example, "Hamilton" could show its ad to a member of the "Broadway Lovers" segment, even if that user is, say, currently browsing the concerts page.

Build segment targeting into your front-end ad product

How you approach this is up to you, but you could offer a targeting category called "Interest Targeting" with a simple dropdown like:

  1. Broadway Lovers
  2. Broadway Purchasers
  3. Current Broadway Attendees

Advertisers (or your internal team) could then create campaigns with relevant ad copy that target just users who fit those segments.

segment targeting example
segment targeting example
Track and store behavior

First, you'll want to create triggers that, when they occur, add the user to specific segments.

Such triggers could include:

  • Specific pages / categories they have visited
  • Web visit frequency and recency
  • Specific products they’ve viewed or purchased

You'll want to then store this information somewhere, and there are two main paths for this:

  • In local storage, such as Window.localStorage. If a user undergoes a trigger action, you could store a string like "segment":"BroadwayLovers" in the localStorage object.

    The data doesn’t leave the browser, is persistent across sessions, and acts roughly like a database. This would be advised over using cookies (which you could employ but aren't great for this type of caching). Be careful, though, localStorage has a seven-day limit on Safari.
  • In a Data Management Platform (DMP), where'd you tie behavior info to a user's row in your database. This would involve a persistent ID, such as a hashed username or mobile ID. You can either build a DMP yourself or integrate with Adzerk's UserDB API.

For instance, your data management platform could have a binary field tied to each user called "BroadwayLovers" that becomes a "1" if the user clicks on a Broadway event.

Make the data actionable

Finally, you need to tie this data with your ad server for targeting.

If data is stored in localStorage, this is where we'd recommend your ad server have a feature we call Keyword Targeting - a way to create rules for targeting keywords attached to each ad request.

Upon each ad request, you would ping localStorage and then pass this information as a keyword in the request. Using Adzerk's targeting tools as an example, you could set up campaigns that'll target only ad requests that contain that segment as a keyword:

interest targeting example
interest targeting example

Or, if the data is stored in a DMP (which we'd recommend), you would need to:

  1. Pass the user's persistent ID in each ad request
  2. Have your ad server cross-reference your DMP in real-time to determine if the user is in a given segment. If it sees that User123's BroadwayLovers field is set to 1, then that user would be eligible to see an ad from any campaign targeting the "Broadway Lovers" audience.

Sound complicated? It is! This is partly why few publishers have added behavioral targeting to their ad platform, even though there’s value. Fortunately for Adzerk customers, behavioral targeting features are built-in and enabling it is easy.

Does behavioral targeting comply with GDPR and CCPA?

Please note: we are not a law firm - we recommend consulting a lawyer for this. Below is for informational purposes only.

Under both laws, persistent IDs are considered PII, and you would need to describe what you are doing in your privacy policy. Where it gets a little grey is that you aren’t sharing this PII directly with advertisers, so there’s no selling/breaching of user data.

That said, to comply with the GDPR, you will most likely need to get consent from EU residents before building user profiles on them.

With the CCPA, though, there is no explicit rule against using PII if you aren’t selling it - meaning you can (likely) allow advertisers to buy against profiles you created using first-party data.

So, you’re saying I should add behavioral targeting to my ad server?

We are! You can generate more revenue with your ad server by offering advertisers more ways to reach their target audiences.



Jane O'Hara

Recommended Articles