Lookalike modeling is a methodology advertisers often use to define consumers most likely to engage with your marketing messages. This model considers common traits or behaviors among current customers, and seeks consumers who share those same characteristics.

Behavioral lookalike modeling differs from other types of lookalike modeling such as demographic or psychographic modeling (that you’re probably already familiar with) in a pretty fundamental way.

Lookalike modeling that uses demographic seeks to find lookalikes who fit into broad categories, like age, sex, income level, marital status and things like that. It’s a very broad and basic way to target an audience who may potentially be in the market for your product or service. So for example, if most of your customers are men between the ages of 25 and 35,  lookalikes based on those customers will match those demographic characteristics.

A step above that would be psychographic targeting.

Psychographics measure and track things like hobbies, interests, habits, and attitudes.

To get an even clearer picture of exactly who we’re marketing to, lookalikes based on psychographics might, for example, look for people who enjoy camping or sailing. If most of your customers share those same interests, it’s a more focused form of targeting and the lookalikes created using that model may be closer to your ideal prospect, but it’s still very general and doesn’t take into account what is the most important determining factor of lookalike creation and that’s behavior.

Unlike demographics or psychographics which can’t tell you what someone is actively in the market for, behavioral modeling tracks things like the keyword searches someone is performing, what videos they’re watching, what they’re doing on social media, the articles they’re reading, the websites they’re visiting and what they’re buying in real time.

What can we tell from that?

Well, when you combine realtime behavior tracking with identity resolution and artificial intelligence, we cannot only know what someone is doing nearly every minute of every day, which includes what they did up to and including the point at which they clicked on a link and landed on your website, but we can also know something else that’s incredibly important.

We can know who that prospect is so we can attach all that realtime behavior we’re collecting to a person. We can know, for example, that it’s a guy named Bob and now once we do that, we have something that’s been the Holy Grail for marketers since the beginning of time:  we know who our prospect is.

We also know what our prospect Bob is doing and has done leading up to the point when he decided to check us out or buy from us.

We also know how to find more people exactly like Bob who are right now doing the exact same things bob did.

We do that through the use of a cryptographic Hash, which is a privacy protected small file that identifies Bob and the things we know about him…..like his email address, and it’s the tool that can be used to find his behavioral lookalikes.

Now what’s great about his Hash is that it never expires. It’s permanent, so once we have it, we can’t lose it. Even if Mom clears the cookies from his computer, the Hash lives on.

The Hash is also portable. If we wanted to reach Bob on an ad platform like Google, Facebook or Twitter, email him, send him a postcard or even reach him on radio or TV, all we have to do is upload that Hash and the platform will know that it’s Bob.

We don’t have to search for Bob and hope we stumble on him by using things like interests or demographics or even the keywords he searching for because we just know it’s Bob, our ideal prospect, which is pretty powerful in itself, but it’s far from the best part.

Because if we have Hashes on Bob and hundreds of other people who did what Bob did, like visiting your website for example, (which means they’re interested in what you sell), we can upload all of those Hashes and then ask those platforms (who know more about what people are doing online in real time than probably anyone else in the world) to go find more Bobs for us, look at Bob’s behavior and find the commonalities and send us more.

That’s how behavioral lookalike targeting works.

It models the most important factor there is, not what people look like or how old they are or what kind of music they like, but it models what they’re doing right now.

So now you know what it is and hopefully can appreciate how powerful it is. How can you make the most of it? Here’s a simple three step strategy to help you leverage behavioral lookalike modeling. First remember that with a Site Visitor Match, you own and control the Hash, which means you own and control all of that behavior that’s represented by that Hash.

So step one, of course, should be to use the Hash as a retargeting audience on the platform that traffic came from.

Let’s say in this case, the traffic came from a Google Adwords campaign. So you would upload the Hash files back to Google and run a retargeting campaign to bring those visitors back to you. But remember, you own and control the Hash.

So as part of step one, upload those Hash files to other platforms and retarget there as well. So if your traffic came from Google, retarget on Facebook, Twitter, and via banner ads for example. Why is this necessary? It’s necessary because it is absolutely vital that you stop thinking of your marketing as a silo type effort and start to see it for what it really is or should be:  which is a seamless and cohesive effort that consistently touches prospects across multiple channels. You should market this way, not because it’s simply a better way to do things. It’s not better. It’s required.

According to Gartner Research, marketing campaigns that integrate four or more channels outperformed campaigns that use only one or two channels by 300 percent. That’s a 300 percent improvement in your campaigns and we’re just on step one.

Here’s step two…… and this is where we really make use of behavioral lookalike modeling because this is where you ask the original ad platform, in this case, Google, to take your Hashes, find lookalikes who matched the behaviors inside those Hashes and give you an entirely new audience, an audience who matches the same behaviors as that initial perfect prospect of Bob.

Once you have that new audience, start a new campaign targeting those new prospects and measure the results. Remember, platforms like Google and Facebook are driven by relevancy, and because you’re targeting the most relevant prospect there is, you’ll be rewarded with higher clickthrough rates and lower costs, but you’re not done……

Since you’re consistently getting a new list of Hashes with new visitors and new behaviors to upload (and recency of behavior is very important when it comes to relevancy) you’ll want to upload those new Hashes and get another behavioral lookalike audience and target them in a new campaign.

Again, measure those results. Do this again and again. Start by uploading new Hash files and getting new behavioral lookalikes every month. Then try every two weeks. Then every week. Test everything to see how it produces and find the sweet spot.

Targeting audiences like this will typically result in a CPC reduction of between 25 percent and 75 percent on average, which translates to a lower cost per acquisition and increased profits, which is the real end game here and what everyone is after.

So step one, commit to a people based multichannel campaign. Statistics say you’ll improve your campaign performance by 300 percent.

Step two, target behavioral lookalikes at the original channel and reduce CPC by 25 percent to 75 percent, but we’re still not done.

Because in step three, we truly take advantage of the portability of the Hash files and the behavior contained in those Hash files.

We upload them to other platforms and ask them for behavioral lookalike using their data.

So if you started on Google, then take the Hashes to Facebook and let their algorithms go to work to give you another audience and so on. Then just like you did in step two, continue uploading new Hash files and modeling the new most recent behavior. Try doing any of that with a platform specific cookie.

These three steps are how you win with Identity Resolution, Site Visitor Match, and the Behavioral Lookalike Targeting………that’s only made possible when you own and control the Hash. That’s how Site Visitor Match does what cookies simply can’t.