Using Lookalike Models to Acquire High-Value Customers, Part 1

This is is the first in a two-part series on the tactics you can use today to acquire more of the customers you want, whether that be by Customer Lifetime Value (CLV), product affinity, price sensitivity, or any other data-driven parameters you can dream up.

But first, let’s take a look at why lookalike modeling is so powerful in an era when acquisition costs are soaring, new competitive sets are emerging daily, and customers are becoming less loyal because of the sheer range of choices and constant marketing inundation.

We’re going to use the example of Customer Lifetime Value because we just love it so much.

Evaluating Acquisitions with Predicted Customer Lifetime Value

Customer Lifetime Value is a prediction of all the value a business will derive from their entire relationship with a customer.

By using CLV to inform more intelligent and efficient media buying, you can move away from proxy metrics like click-through rates and cost-per-action to a more accurate calculation of return on investment.

The idea here is to tweak the paradigm of “more customers for less acquisition spend” to “more revenue for less acquisition spend.”

Here’s a simple example.

_190606_BLOG_UsingLookalikeModels1_Embedded1

Let’s say you’re serving an ad for your cat t-shirts to two Facebook users, Sonya and David.

David’s cost per acquisition (CPA) is $9, while Sonya costs $6 to acquire.

Sonya may have been cheaper to acquire, but she only buys one $9 shirt and never shops with you again.

David, on the other hand, continues to purchase from your business. Over the course of his life with your brand, he will buy $36 worth of cat t-shirts.

So while David might be 33% more expensive to acquire than Sonya, he purchases 400% more and generates 6x greater ROI than Sonya.

In the long run, customers like David are more valuable to your business, even though they cost more to acquire.

Now the question is, how do you use these insights to improve the performance of your acquisition campaigns?

That’s what we’re thinking too. So let’s get into the best way to prepare for your most awesome lookalike campaign yet.

The First Step is First-Party Data

The first step in acquiring new high-CLV customers involves identifying a seed audience—a segment of existing high-CLV customers whose key characteristics can be used to home in on similar people in the general consumer pool.

First-party data about customer interactions with your brand will give you much more accurate information about your customer base. Some of this data includes:

  • Website behavior (number of visits, minutes spent per session, products browsed)
  • Purchasing history (products purchased, purchase amounts, coupons used)
  • Email engagement (emails opened, clicked-through, and sent)
  • Call center recordings
  • And more!

Your first-party data can be incredibly wide-ranging and can give you a comprehensive view of your existing customers.

Provided you have adequate data collection and organizational processes in place, you can—and in most cases should—build your seed audiences exclusively around first-party data.

Among other benefits, doing so guarantees your lookalike prospecting efforts will be aimed at an audience of people who are reasonably likely to have a significant affinity for the brand.
 


Cave Venditor (Seller Beware): Common Seed Mistakes

You can put all your effort into growing a mighty oak tree, but if you planted a fig tree, you’re figged.

Mistake #1: Using Demographic Data

There are as many ways to segment the customers in your database as there are customers in your database. Different use cases call for different segmenting criteria.

Your creative team would certainly benefit from traditional qualitative persona research that provides demographic information on age, gender, location, and a range of stated preferences, like favorite online and offline publications, TV shows, clothing brands, etc.

But these descriptive elements do nothing to inform you about how someone interacts with your brand.

For instance, perhaps Jill and Jane live in apartments across the hall from each other. They are the same age, race, and gender. They are both sales execs at the same company and earn the same amount. They are both single and own cats (Roastbeef and Ray, respectively). But they couldn’t be more different in their shopping preferences and behaviors.

If you want to acquire customers that behave in a certain way, you need to build your seed audience around behaviors, not demographics or stated preferences.

In fact, when we say “lookalike,” just go ahead and mentally replace it with “do-alike.” That’s what you’re looking for—a do-alike audience, hence the behavior-based seed.

Mistake #2: Using Third-Party Data

Your ability to effectively identify a seed audience hinges on the sophistication of your data infrastructure. Without mature data collection and organization protocols in place, you’ll be forced to rely on third-party datasets, which can be problematic for a variety of reasons.

First and foremost, third-party data describes actions taken by a specific web browser rather than a specific customer. This browsing activity is aggregated by a third-party vendor and only then sold to marketers. That means you’re at least two steps removed from the digital activity about which you’re buying information.

Combined with the data’s anonymization, this disconnect from the real, flesh-and-blood people described by third-party data prevents you from tracing and adjusting for the logical jumps that often underlie these datasets.

For instance, just because a particular device is used to watch a string of Netflix original rom-coms doesn’t mean the device’s owner will be interested in purchasing bouquets and bonbons, but these are exactly the type of tenuous connections that third-party data vendors tend to draw on when assembling audiences to sell to a florist or chocolatier.

Step 2: Find Your Gold Reserves

Once you’ve spun your first-party data into a seed audience populated by your highest-value customers, your next step is to run this audience through a lookalike model to find a larger audience that reflects the benchmark characteristics of the seed group.

Facebook’s Custom Audiences tool lets you upload your seed list, based upon which the tech giant’s proprietary algorithms generate a lookalike prospect pool. In the social space, this is typically accomplished by sorting users based on demographic information and their historical web activity and engagement.

Google Customer Match works similarly. Retailers upload a seed list to AdWords, and the search engine automatically optimizes bidding through enhanced cost-per-click metrics. Customer Match empowers retailers to fight harder to win bids for ad space that appears before their highest potential CLV audiences, giving them the option of shelling out a slightly higher CPA for a customer who’s likely to cover the premium.

Finally, identity-resolution tools like LiveRamp allow retailers to move beyond walled gardens like Facebook to conduct cross-channel lookalike prospecting on the open web. Once you upload a seed list, LiveRamp matches potential high-CLV customers with the targetable cookie IDs that are applicable across a variety of demand-side platforms (DSPs) and data management platforms. You can then plug these targetable IDs directly into your adtech tools, enabling you to follow prospects across various marketing channels.

In the next post in this series, we’ll get into how to prioritize channels by business challenge, how to improve your seed list to improve your lookalike audience, and how to use predictive analytics to get granular with your targeting and to optimize each campaign with a surfeit of first-party data.

In the meantime, you can learn more about the absolute most important customer-centric KPI (spoiler alert) in our latest book, The Chance of a Lifetime: How to Use Customer Lifetime Value Reporting to Grow Your Retail Business.

 

Like this? You might also enjoy these.

How Supergoop Topped Custora’s Retail Benchmarks, part 1

This past May, Custora announced a new benchmarking feature within our...
Read

Your Revenue Forecasting Ignores the Source of Your Revenue

If you’re like most organizations, your finance department focuses on overall...
Read

4 Steps to Solving the One-Time Buyer Problem, Part III

In parts one and two of this series, we laid the groundwork on a methodology to...
Read