For a long time, advertisers have relied on traditional behavioral targeting to reach new customers. As marketing technologies improve, more businesses are turning to predictive audience targeting, fueled by customer intent data, to drive higher performance by more accurately identifying users that are most likely to convert.

The Limitations of Behavioral Targeting

Traditional behavioral targeting uses high-level information about a person’s characteristics and expressed interests to deliver ads based on static or demographic data. Marketers typically set rules using properties they think define their customers. For example, “If a user lives in X city and is within X age group, serve them X ad”.

While behavioral targeting can get your brand in front of audiences who might be interested in your product, it doesn’t consider the full picture of all the diverse factors that affect a purchase decision. People are much more dynamic than the demographic groups they might fall into, and individuals that “look” the same may not necessarily act as such.

Take, for example, two 30-year-old male Portlandians who both enjoy spending time outdoors, and are partial to flannel shirts. Behavioral targeting aimed at this demographic would assume that they have shared interests, when they’re actually in-market for different products at different times. Acting on static or demographic data alone is always just a guess at what a user might be looking for.

customer profiles 2

Perhaps you’re a men’s retailer targeting these two individuals based on their demographic. In reality, neither of them are looking to buy new clothing, and aren’t interested in your ads. Meanwhile, there’s a girlfriend you didn’t consider that’s been shopping around for new sweaters for her boyfriend. She’s in-market and would be highly likely to engage with your brand. Traditional behavioral targeting would overlook this potential customer.

So, what’s the alternative?

How Predictive Audience Targeting Works

Previously we’ve talked about intent data, which lets you know exactly when someone is in-market for your products. Intent data captures the customer intent signals a user sends out while browsing and shopping online. This is the data that powers search campaigns and dynamic creative ads.

According to eMarketer, 71% of advertisers already use real-time, self learning analytics to drive personalization and targeting across digital channels, while another 57% employ machine learning and artificial intelligence algorithms to personalize interactions. Advertisers use intent signals to drive results by targeting those users who are most likely to convert, with customized ads that resonate.


The best predictive audience targeting tools layer on powerful algorithms over a strong dataset to map out customer intent signals. Most advertisers have already started to adopt these tools with regards to their own first-party data set – using retargeting, for example, to leverage the customer intent data users generate when browsing their website.

Now, imagine if marketers could access an even larger pool of intent data that helped them understand what their customers are doing before and after reaching their site. This is where the concept of pooled advertiser data arises.

A lot of users display similar behaviors when shopping for certain products – researching complementary items, comparing related purchases, or visiting specific review sites or recommendation engines.

Predictive targeting tools that act on pooled datasets capturing these various behaviors can be exponentially powerful for businesses. They allow advertisers to make use of additional insights on their customers, outside of what they know from interactions taken on their own site. Marketers can target new users who display similar behaviors to existing customers, and are therefore much more likely to convert than someone who hasn’t displayed any of these behaviors signaling intent.

Pooling Intent Data to Drive New Customers

AdRoll’s Prospecting solution, currently in Beta, works in this way. We’ve spent a lot of time developing this product to allow advertisers to take advantage of the massive amounts of data we’re getting on how, when, and where customers shop.

prospecting1Here’s how it works: an advertiser can opt-in to the Prospecting data pool, allowing us to utilize their first-party customer intent data in our predictive engine for prospecting. What they gain is access to a powerful act-alike algorithm powered by dynamic intent data from over 1500 advertisers. Marketers can run campaigns that reach new audiences, with targeted ads based on real-time interest. Prospecting campaigns use the same bidding algorithms that fuel our world-leading Retargeting product.

As our opt-in advertiser pool grows, so does the power of our predictive engine and our ability to identify new in-market prospects for a particular brand or product. We’ve collected a wealth of information on users’ purchase intent, and are continually improving on the way in which we act on this data to help advertisers grow their businesses.

Early performance indicates that CPAs drop by 50% – 80% when moving from behavioral targeting to predictive behavioral targeting via Prospecting. These campaigns fuel even better results for Retargeting.

Retargeting has proven that personalized advertising driven by customer intent data delivers strong performance for marketers. Now, advertisers are looking for more data-driven solutions, based on demonstrated user intent, to find new customers.

Predictive audience targeting is more powerful and precise than behavioral targeting, utilizing intent signals to drive higher performance and more, higher quality net new customers.
Check out AdRoll Prospecting to learn more about how predictive audience targeting works to help marketers engage with customers.