In-depth guide to Facebook Lookalike Audience Ads for e-commerce stores

A proven formula to help you drive quality traffic to your store

Reza Khadjavi
Shoelace
11 min readNov 6, 2017

--

Many of the most advanced e-commerce marketers we work with at Shoelace have built their entire top-of-the-funnel Facebook ads strategy around Lookalike audiences. It is an extremely powerful way to reach new prospects who are similar to your existing customers.

In this post, we’ll be covering the fundamentals of Lookalike audiences and how e-commerce advertisers can use it to drive high quality traffic to their stores. We’ll also be talking about the “Conversion” campaign objective and how combining Lookalike audience targeting with this campaign objective in Facebook can be a winning formula for your prospecting ads!

Lookalike Audiences + Conversion Objective = 🔥🔥🔥

What are Lookalike audiences?

In 2013, Facebook first launched Lookalike audiences to allow advertisers to target prospects who are similar to their existing customers. The wealth of data that Facebook collects on it’s users has always been what sets it apart from other ad networks and what makes it a special place for marketers to reach new customers.

Not only does Facebook collect user activity data on Facebook and Instagram (who their friends are, what pages they like, posts they engage with, etc), but it also collects data on what it’s users do across the Internet. For example, any website that is running a Facebook pixel, a like button or a “Sign up with Facebook” login system, is also sending data back to Facebook about actions users perform while on that site or app. This gives Facebook a massive amount of data on each of it’s 2 billion users!

This is what allows Facebook to implement something like Lookalike audiences so successfully. When you create a Lookalike audience, you first upload a “source audience”, this can be an email list of every customer who ever bought from your store (more on this below). Facebook crunches all the data it has stored about these users to find similarities among this group. The similarities are not just restricted to things like age or interests. A lot of what Facebook looks at are behavioural components too. What are the shopping habits of these users, where do they buy from, how frequently, etc, etc.

Then, Facebook’s algorithm will start to query it’s entire user base looking for people who most closely share the profile and behavioral traits that it has detected from your source audience. The result is a Lookalike audience of prospects that are very similar to your existing customers.

Setting up Lookalike Audiences

In your Facebook Ads Manager, under the Audience section, there is a button called “Create Audience” which allows you to create a Custom, Lookalike or Saved audience.

Once you select Lookalike Audience, you will see a screen like the one below which requires a Source, Location and Audience Size in order to create the audience.

1. Setting the Source

Source is the existing audience that you want Facebook to use as a base in order to find a Lookalike audience of similar people. You can use an email list, a custom audience from your Facebook Pixel or even just fans of your Facebook page.

For a simple first approach, here’s what I recommend:

Gather the email address of every customer who’s ever purchased from your store and use them as the source audience.

If you’re not sure how to set that up, under Source click the Create new button then choose Custom Audience > Customer File > Add customers from your own file, then proceed to either upload your CSV file or just copy and paste your email addresses directly into the box.

Minimum audience size for Source

The recommended source audience size should be at least 1,000 people or more. The bigger the audience, the better Facebook’s algorithm is going to work for finding a strong Lookalike match. For best results, the quality of your source audience really matters, too. A source audience made up of your most loyal repeat-purchase customers will be much better than a source audience of people who just bought from you once, for example.

What if you don’t have 1,000 past purchasers?

If you don’t have a list size of 1,000 people who have previously bought from your store, here are a few other sources you can use instead:

  1. Only have 100–200 past purchasers? Just use them anyway and screw the 1,000 minimum! (your Lookalike audience just won’t be as strong)
  2. Have an email subscriber list? Use that as the source audience (ideally use the most engaged subscribers e.g. the ones who open/click your emails).
  3. Have some website visitors? Create a custom audience of people who visited your website (ideally use the audience of people who added to cart or at least viewed a product).
  4. Have some fans on Facebook? Use your Facebook fans as your source audience to create a lookalike.

2. Choosing the Location

If you have one major market that you sell to (e.g. United States), choose that. If your store serves multiple markets, my advice would be to create a separate Lookalike audience (using the same Source audience) for each of the geographies you want to reach. That way, you can run separate ads testing the various locations to see how they each perform.

3. Picking the Audience Size percentage

The audience percentage selection here simply means how big of an audience size do you want Facebook to find, as a percentage of the total population you’ve selected. For example, if you choose United States and 10%, you’ll get back an audience size of roughly 21M people.

For a simple first approach, here’s what I recommend:

Start with 1%. These users will be the most closely matched with your source audience.

The common practice from the marketing pros is usually to start with 1% and once they feel like they’ve scaled as far as they could with that audience, they start to broaden out to 2%, 3%, etc. But it will take 10s of thousands of dollars in ad spend (or more) before that starts to happen, so no rush.

Setting up the Ads

Now that you have your Lookalike audience set up, it’s time to create some ads! The winning formula we’ve seen e-commerce marketers use the most, combines Lookalike audiences with Conversion objective ads.

It is easy to overlook this step as just Facebook’s way of trying to understand what your marketing objective is, but what you pick here actually matters a lot. If you’re an e-commerce brand and your goal is to drive sales, you’ll want to select the Conversions objective here.

What’s so special about the Conversions objective?

When you select Conversions objective, Facebook will let you pick a Conversion event on the Ad Set level to optimize for. Here, you’ll pick the Purchase event.

This is not just Facebook’s way of trying to learn what your marketing goals are. What you pick really does matter. Earlier, we explained how Facebook collects a ton of rich data on it’s users online activities. One of the activities that Facebook monitors closely is when users make purchases online.

Whenever a user makes a purchase online, Facebook logs that event to the user’s profile and uses that to determine which of it’s users are most likely to shop online.

By selecting the Purchase event as the conversion to optimize for, Facebook will crunch it’s data to isolate the users in your audience who are most likely to purchase from an online store. But how will it know? Whenever a user makes a purchase online, Facebook logs that event to the user’s profile and uses that to determine which of it’s users are most likely to shop online.

Optimize for conversion window (1 day vs. 7 day)

The other benefit of selecting the Conversion objective is that it gives you the option to select a conversion window to optimize for.

I recommend starting off with the “Conversions” , “1 day click” and “Automatic bid” settings here.

Facebook records data on every one of its users from all kinds of different data points. One of those data points is the purchases that users make online after clicking or viewing an ad. If a user clicks on an ad and converts within 1 day, Facebook logs that activity on that particular user. If another user clicks an ad and converts within 7 days, Facebook will record that too and log it on the data it stores for that user.

When you select a setting like this, you are asking Facebook to prioritize the delivery of your ads to people who in previous interactions online, have clicked on an ad and converted to buy within 1 day (or whatever you’ve picked).

Facebook has a very clever way to prioritize ad delivery to those who have performed actions of the kind you’ve specified here, in hopes that they’d behave similarly once seeing your ads (as they have demonstrated such behaviour before).

Naturally you’ll want to target people who will click your ad and buy within 1 day of that event. So I often see the marketers starting with the 1 day click configuration here and expanding out to broader windows once their audiences have started to saturate. However, as in the Lookalike audience example, it can take 10s of thousands of dollars before the audience starts to saturate. So for most people, 1 day click can go a very long way.

Detailed Targeting Setting

In every Facebook Ad, you’ll have the option to set detailed targeting options like Interests, demographics or behaviors.

My most common piece of advice here is to leave this section empty. The combination of Lookalike audiences and Purchase Conversions optimization is usually enough to let Facebook do it’s magic to narrow down the ideal audience to deliver the ads to.

The combination of a strong Lookalike audience and Purchase Conversion objective is usually enough to let Facebook do it’s magic to narrow down the ideal audience to deliver the ads to.

Choosing the right Ad Creative

This is the part you’ll want to experiment with the most. But a common set up that I see a lot is a carousel ad featuring engaging lifestyle images of some the brands most popular products.

However, you don’t want to make too many edits to your ads too frequently. Here are a couple of reasons why:

  1. Your ads will start to get likes, comments and shares. The positive engagement on the ads will act as social proof to get visitors to pay attention as they scroll through their feed.
  2. When you make an edit to the ad, your previous post engagements will get erased and you’ll have to start from scratch.

Determining the right budget

The biggest mistake I see new advertisers make is spending way too little and giving up way too early. The key thing to remember is that Facebook’s algorithm gets smarter over time. If you follow the steps above, you’ll have given Facebook a ton of data to work with in order to show your ads to the right people. However, a very important part of Facebook’s ad delivery algorithm is learning from the data it collects from the campaign itself. And it costs money to let a campaign run, so you’ll have to be comfortable spending some money towards this.

You should be willing to spend roughly $50 to $100 per day for about 1 week before calling it quits on a campaign.

I’ve seen advertisers using more or less the formula I’ve described above, spending $10K to $20K per day on ads! $50 to $100 per day is probably the minimum you should be willing to spend (for about 1 week) before calling it quits on a campaign.

What kind of ROAS to expect

In general, if you are seeing 1x to 3x return on ad spend, your campaign is doing well. That means if you spend $700 in ads over the course of a week you’ll want to see somewhere between $700 to $2,100 worth of sales as a result of your ads. That is more or less the range that successful advertisers are seeing with this sort of campaign.

The impact of average order value on ROAS

Average order value is revenue divided by number of orders.

Average order value (AOV) is a very important metric that plays a huge role in your ultimate ROAS. It’s often the difference between a successful or failed campaign.

Let’s look at two scenarios to see the impact AOV has.

In both scenarios, you’ve spent $500 in ad spend and got 23 purchases. But in one case that resulted in $1,495 in sales and in the other it was $575 in sales. That’s almost 3x ROAS in one case and 1x ROAS in the other. Can you see how the AOV can make all the difference?

To increase your likelihood of success with Facebook ads, you should have an AOV of $50 or more.

I’ve seen advertisers with low AOVs still find a way to achieve success, but for the most part, if you want to increase your likelihood of success with Facebook ads, you should have an AOV of $50 or more.

Conclusion

It can be easy to get overwhelmed with all the different types of ads and experiments you can be running to drive traffic to your site. When you’re just getting started, the best approach is to start off simple and add more complexity as you go.

I hope this guide helps you lay the foundation of your Facebook ads strategy. Driving high quality traffic to your store is essential, but once you start driving traffic, you’ll want to have a solid retargeting strategy in place as well. On average, 98% of visitors will leave without completing a purchase. Retargeting allows you to get in front of those users again and encourage them to come back. Sign up for our newsletter below to be notified when we write about Facebook ads, retargeting or e-commerce marketing in general.

Happy retargeting!

-Reza Khadjavi
CEO at Shoelace

If you want an intelligent retargeting expert to take care of all your retargeting for as little as $49/month, consider trying Shoelace today!

--

--