CategoryShow

Advertise some products to audiences who are similar to the current customers or a specific segment of people

Targeting a close audience to your current customers might be a profitable strategy. Ad networks like Google, Facebook, or LinkedIn have a lot of data and powerful algorithms you can take advantage to detect similar behaviors between their users.

What is the process?

Create a similar/lookalike audience on Google, Facebook/Instagram or LinkedIn based on your current customers (or a specific segment).
Set up your conversion trackers and goals to measure the true impact of your campaigns.
Analyze your CPA and the profitability you get from your ads. Remove the inefficient ones, increase the audience and optimize what works.

Why should you care?

Thanks to their enormous amount of data and their machine learning expertise, Google and Facebook have designed powerful algorithms to detect similar behaviors between their users. Targeting an audience close to your customers (or who belong to any important segment) is among the first campaigns to try (after retargeting) to make the most out of your budget.

Facebook and Google know what they are doing with their algorithms. We targeted similar/lookalike audiences for some e-commerces in the past, and we succeeded at getting some return on investment (not on every campaign though). Even if we had to give up some control over who should be in the audience, we had a few quick wins.

If you do not want to play with all the standard interest-based audiences that these ad networks provide, similar/lookalike audiences can save you a lot of time while still giving you profitable results.


What tools will you need?

Unlock all our Shopify experiments
This content is restricted to our subscribers but you can read our last 10 experiments for free.
You will also get access to all the features on the platform without any restriction.

How to execute this experiment?

1

To target people similar to your site’s visitors on Google Ads (i.e., Search Network, Display Network, YouTube, and Gmail), you will have to create a similar audience. A similar audience is a targeting feature based on first-party data lists (most commonly remarketing lists), that help you expand the reach of your best-performing audiences by targeting new users with similar characteristics. Once set up, Google will update them automatically in real-time using its machine learning capabilities. The seed list must also include at least 100 people, and it can take up to 48 hours to generate the new audience.

2

On Facebook, it is called a lookalike audience. A Lookalike Audience is a way to reach new people (on Facebook, Instagram and Messenger) who are likely to be interested in your business because they're similar to your best existing customers. Your seed data must that will serve to generate the audience must include at least 100 people who have a valid Facebook account. You should also start with a 1% audience when you have to choose the size since it will increase the level of similarity (relevance). Then, increment it one percent at a time until it does not bring any profits. Depending on the country(ies)/region(s) you have selected, 1% can still reach a broad audience. For instance, 1% in the United States will roughly target 2 million people. It takes between 6 to 24 hours to generate the lookalike audience. Facebook refreshes it automatically every 3 to 7 days as long as you are still using it in one of your campaigns.

3

You can also target people with similar characteristics to your source audience on LinkedIn (which will be identified using an existing matched audience segment, such as website audiences, account lists, or contact lists). You can also refine a lookalike audience by adding professional targeting options. This professional network can be useful when selling B2B products to decision-makers. Just keep in mind that the CPC is usually higher than on Facebook and Google.


How should you track?

Head over the ad manager of each social network you advertise on to assess how your campaigns are performing. Make sure to set up goals and conversions so you can measure your cost per purchase (CPP) and cost per activation (CPA).

Google Ads provides different ways to track conversions. In most cases, you will be interested in using the conversion tracking pixel to track sales and other essential actions on your website (signups, form submissions, etc.). The link below will give you what you need to know to set it up. Don’t forget to set a value for your conversion. Especially if you are selling products at different prices. It will help you calculate things like your return on ad spend (ROAS).

For Facebook and Instagram, you will have to set up your ad objectives.

The process is not different for LinkedIn as you can set up conversion tracking to measure the true impact of your ads.


Key learnings.

A high CTR on a specific ad means that you found the right message for the target. If your product page is converting well after that, this is usually the path to significant campaign profitability. Also, keep in mind that high CTR correlates with low CPC and high-quality score (which is rational as your content is relevant to the audience, so the network is billing you less for each click).

Make sure you are building your similar/lookalike audiences on a list that includes at least 1,000 people. You are making things harder for your ad network if it does not have enough data to find in its users, people with the same characteristics.

All customers are not equal. You can segment your list further by grouping them according to their average order value, lifetime value, or by the product category they purchased. For instance, this might be a good idea to generate a lookalike audience based on a small list of your customers who make most of your revenue. This tactic might allow you to bid higher and still be profitable

If you are in a situation where the ad network does not have a reasonable rate to match your customers, you can still target all the people who visited your thank-you page (if your pixel is in place). For instance, this can be useful if you only have B2B emails as they have a lower match rate than B2C.

The most relevant your seed data is (in quantity and in quality), the better results you will get. Facebook recommends using a source audience that contains 1,000 to 50,000 people.

The interesting source audiences you can use to generate your lookalike/similar audiences are your current or most valuable customers, your website visitors (or those who viewed a specific product page), your past purchasers and people who added something in their cart.

To make sure you only target new users, Google, Facebook, and Linkedin exclude the source list from the generated lookalike/similar audience. The only situation where Facebook does not exclude them is when you use a pixel as your source audience.

For Google Ads, the generated audience does not apply any geolocation restriction. You will have to select the countries/regions in your campaigns to limit the coverage to a specific geographical area.

If you do not have enough reach: increase your bid (sometimes a low bid only gains you access to low-quality and small inventory) or use a source audience that is bigger.

If Instagram plays a significant role in your marketing strategy, you can build a custom audience of people who are engaging with your profile and then create a lookalike audience from it.

Once again, knowing the lifetime value and average order value of your customers is everything to know what moves you can or should not do.

You can now create a value-based lookalike audience on Facebook. It will take your customer lifetime value into account to help you better reach the people that matter to your business.


A comment, an improvement, an issue?

Engage a conversation with us.