4 Challenges Marketers Face Moving from Segmentation to (More Effective) Lookalike Audiences
What’s a lookalike audience and how does lookalike marketing work?
Lookalike marketing is a popular practice of using existing customer profiles to find similar potential customers to market to. The concept is based on the assumption that a group of people with shared attributes and behaviors will respond similarly to certain offers or campaigns.
Digital marketers are always looking for more effective ways to attract visitors to their products or services. It turns out that one of the easiest ways to accomplish this is by identifying prospects who look like existing customers — and then targeting them with offers.
This technique has significant advantages over more traditional strategies of targeting offers based on data such as demographics or geographic location. (Not all men aged 18 to 25 like the same things; not all people from Houston like the same things).
Say you’re an outdoor retailer hoping to target people who are likely to buy your products. Lookalike marketing would be an efficient and cost-effective way to find your target audience. It begins by using machine learning to find more users who are likely to purchase your products — perhaps people who shop with your competitors or people who have searched for products that you sell. Your lookalike marketing campaigns will return a higher engagement rate than ones that targeted specific geographic or demographic data. How much higher? According to one source, lookalike audiences average a higher click-through rate than other audiences 90% of the time.
In addition to higher conversion rates, another advantage to lookalike marketing is that it helps marketers identify a larger pool of possible qualified customers. So expanding the pool in a targeted fashion, rather than “spray and pray” advertising tactics. With lookalike marketing, you identify audiences with behaviors that match your target audience, giving you a better chance to convert them to buyers. It can also be used to find other similar customers who fit the same profile on many digital platforms which provide powerful tools to support lookalike marketing, such as Google, Quora, or Facebook.
The 4 Hurdles to Implement a Robust Lookalike Marketing Program
You may be thinking that marketers have always relied on some sort of version of lookalike marketing — identifying customer profiles and then aligning marketing programs to fit that profile. There is some truth to that, but today’s Lookalikes are much richer. Before marketers can get the full benefits of a lookalike marketing program, there are some key hurdles marketers need to overcome. We’ll discuss four of these hurdles below.
1) Big Data Enables Better Creation of Lookalike Audiences, But It’s Not Easy
A key thing that is new is the sheer amount of data currently in the hands of marketers. The quantity of data—already overwhelming — is increasing exponentially every year. In fact, 90 percent of all of the world’s collective data was created in the last two years, with an additional 2.5 quintillion bytes generated each day. Gathering and interpreting that much data is hard, but bringing it together to build a 360-degree profile is even harder. This data can include any or all of the following:
- Customer Record (name or customer ID)
- Customer Accounts
- Orders / Purchases
- Services / Repairs
- Customer Support Records
- Feedback / Reviews
- Payment Methods
- CMS / Marketing Offers
- Marketing Sources
- Survey Data
- Delivery / Shipping Data
- Loyalty Program Data
2) GDPR Complicates Data Collection
The European Union’s General Data Protection Regulation (GDPR), which took effect in May 2018, has made it more complicated for companies who buy and use third-party data to stay compliant, forcing them to do more due diligence rather than simply trusting their data broker’s word. The need for scrutiny may eventually increase transparency, but in the meantime just makes things more difficult. With the right platform, today’s marketers can identify similar prospective customers on online platforms — for example, Facebook and Twitter — rather than having to buy third-party data to market to. For more information, download our Marketer’s Guide to GDPR.
3) More Data Sources Provides an Opportunity for Richer Customer Profiles
Today’s customer profile can include an enormous amount of granular data and details on web and mobile behavior. Remember when the most we knew about our prospective customers was that they were females between the ages of 21 and 30? The challenge of moving from segmentation to lookalike marketing is the necessity for a system that pulls these disparate pieces of information together. A customer data platform (CDP) can accomplish this by storing all of the data in one place.
Now, with CDPs, we have the opportunity to create complete profiles gathered from customer service data, customer data, CRM data, etc. With rich information on purchase behaviors and buying preferences, combined with third party demographic data, we start from a much better place to identify an even more targeted segment of prospective customers.
4) Machine Learning Powers Creation of Lookalike Audiences
Another hurdle is the ability of any one system to analyze and make decisions based on all of the different pieces of data. There is simply too much data for human brainpower to efficiently process. To solve this challenge, marketers will need to tap into the power of machine learning (ML) and artificial intelligence (AI). These technologies have the power to digest and parse through much more data to continuously refine these customer profiles. Even with thousands of attributes in the system, machine learning can begin to filter in on the variables that contribute to conversion in a meaningful way. Behaviors with a high propensity to convert are typically very specific, occurring with low frequency — making it much more difficult and time-consuming to create good models without machine learning.
How to Get Started (Hint: You’ll Want a CDP)
To fully reap the benefits of lookalike marketing, you will need extensive sources of data about your customers. A good CDP allows you to integrate your customer data sources into one central repository, where you can access it in one place. This single source of customer data can encompass both online and offline engagement, and can all of your customer touchpoints, including website behavior, email engagement, purchases, loyalty programs, customer support tickets, product reviews and more. The result is a much richer and more accurate view of customer behavior.
Treasure Data enterprise CDP offers integrations with hundreds of different data sources, including Unity, Google Analytics, Instagram, Salesforce, Marketo, Shopify, Zendesk. The combination of the online and offline data is powerful enough to set you apart. And if you’re looking for additional information on how to capture offline data, read our loyalty program blog post for inspiration.
With your data centrally stored in one place, the first step in lookalike modeling is to identify people who look and act like your target audience. This process is greatly simplified by using a tool — such as a CDP — that can analyze your seed audience, identify their key attributes, and then look for similar customers. A CDP allows you to leverage machine learning (ML) and artificial intelligence (AI) to analyze all of this data.
Once your audience (or audiences) are defined, your next step is to test these target segments on online platforms which allow you to quickly understand if you’ve found the right profile. You can choose from a number of Treasure Data’s integrated services to upload your targeted list to. Select services from the appropriate menu and with just a few clicks, you can send your target list to all the platforms you need for your multichannel campaign.
Results from Better Segmentation and Targeting
What can you expect to see from better segmentation and targeting for your campaigns?
For one thing, you can expect improved advertising effectiveness through this timely and personalized marketing.
Panasonic is a Japanese multinational electronics company with four different companies valued at over $20B apiece. Its marketing model has always relied heavily on advertising to drive sales. To ensure the success of their digital marketing efforts, the Panasonic marketing team needed a way to leverage all of the customer and visitor data being captured through fan websites, product support sites and 3rd party advertising.
Panasonic built a team of specialized data marketers —The Data Marketing Center — to conduct sophisticated analysis for its marketing team, integrating data from multiple systems, such as eCommerce websites, competitor websites and 3rd party product comparison sites. With such an enormous scale, Treasure Data was the Panasonic team’s choice of solutions for the platform. The Data Marketing Center created a multi-dimensional customer model that could be used for more than 350 segmentation requests at a time (35 internal marketers x 10 promotions each).
The ROI in Your CDP
Lookalike marketing programs can be highly profitable, leading to improved conversion rates and more targeted marketing opportunities. But there are several hurdles including the number of data sources that need to be integrated, the massive volume of data, managing GDPR responsibilities, and the need for ML/AI to assist with analysis and decision-making.
A CDP is an essential element of solving these challenges. To learn more about the various advantages of an enterprise CDP, download our new whitepaper, “The ROI in a Customer Data Platform.” It details how marketing organizations in various industries are generating revenue growth through an enterprise CDP.