How To Choose Audiences For Top Of Funnel

We spend a lot of time building out COLD Traffic (i.e. Top of Funnel campaigns). Our top of the funnel campaigns usually spends around 65-70% of the total budget on most accounts we manage. So, today we're diving into how to choose the right audiences for your TOF (aka Cold Traffic) campaigns.

Share This Post

Share on facebook
Share on linkedin
Share on twitter
Share on email

Cold traffic audiences are, without a doubt, the most difficult audiences to profitably convert with paid advertising. However, they’re usually the most important audiences you run ads to, even if your bottom of funnel audiences provide you with the greatest returns. If you’re not bringing new people into your funnel and building your pipeline, then your pool of potential customers at the bottom of the funnel will have diminishing returns.

We spend a lot of time building COLD traffic (i.e. Top of Funnel campaigns). Our Top of Funnel (TOF) campaigns usually spend 65-70% of the total budget on most accounts we manage.

Because so much ad spend is directed to cold traffic, it’s essential to choose the right audiences for your TOF campaigns.

Traffic quality is vital to your ad account’s overall performance. Since this traffic feeds into the Middle Of Funnel (MOF) and Bottom Of Funnel (BOF) campaigns, having the right audiences at the top will give your account the opportunity to perform well.


Tired Your Campaigns Being Unprofitable?

Learn the method we’ve used to help profitably grow and scale e-commerce brands to 6-7 figures online with paid advertising. Save time, money, and countless hours testing by following our scaling method.


To set up your top of funnel, evaluate your options and separate the available assets into 4 main categories:

1. Lookalike audiences (LLA) based on Custom Lists.

Custom Lists are not new. If your Facebook Pixel did not capture purchases, upload a CSV and create a lookalike audience.

PRO TIP: Export your customer list from your ecommerce platform (Shopify, BigCommerce, WooCommerce, etc.) and separate your customer data based on the amount of purchases they’ve made. Our favorite starting point is to segment people that have made only one purchase vs the people that have made 2+. We then take this data and upload the lists separately, then our final upload is both of those customer lists combined. This provides three solid audiences for testing and creating Lookalike Audiences from 1%-6%. Make sure not to remove the order value from the list so you can upload the Custom List with Life Time Value included.

Customer List #1 = People that have made only 1 purchase

Customer List #2 = People that have made more than 1 purchase

Customer List #3 = Everyone that has made a purchase

2. Lookalike audiences based on pixel events.

There are countless options and variations when creating lookalike audiences based on pixel events. The most common and available pixel events in ecommerce are PageView, ViewContent, Add to Cart, Initiate Checkout, Added Payment details, and Purchases. Before creating custom audiences based on these events, you’ll always want to check Events Manager for the total number of events fired on your website for each event. These numbers can help predict the potential audience size and whether to use them as lookalikes or not.

Regardless of the total number of events, we recommend creating custom audiences for 3 / 7 / 15 / 30 / 60 / 90 / 180 days. Even if you cannot use them as a source audience for lookalikes, they will come in handy for retargeting later on.

If the number of events on your website is significant or the site has several products or landing pages, it will allow you to create more advanced custom audiences, providing more valuable lookalike audiences. We aim for the custom audience size to be at least 500 (preferably 1000) to create a solid lookalike audience. 

Here are some of our favorite LLA combinations:

  • PageView Last 30 Days Frequency > 4
  • Site Visitor Top 5% Last 90 Days
  • Top 25% Most Engaged By Time On Site
  • Any specific product / category / pixel event –> View Content_url include: Product last 60 days
  • You can also exclude purchase value under a certain amount

PRO TIP: To do an advanced search or crosscheck the potential custom audience size, you can use Facebook Analytics. It will immediately show you the audience number, and you can adjust variables more easily than creating each custom audience in the Audiences tab. If you find a great combination, you can save the filter and build the audience right away.

3. Lookalike audiences based on Engagement.

This category could be called “lookalikes based on MOF audiences”. The source audience, in this case, is from custom audiences based on Facebook Page or Instagram account engagement.

Here are a few examples:

  • FB messages sent in the past 365 days –> LLA for a client where most of their orders came through Messenger
  • FB post engagement in the past 180 days
  • FB video watch time 50%+ last 90 days
  • IG saved post in the past 180 days
  • Facebook Page visitors in the past 60 days
  • FB all Engagement in the past 365 days

Note: We believe you should only use video viewers as a source audience for lookalikes if there is no other option available. The source audience that viewed your ad videos aren’t necessarily valuable customers. Creating a lookalike from these users might not provide much value — unless your objective is to find more engaging video viewers. 

There is, however, one scenario when video viewers can be valuable as a source audience: when you have organic converting traffic with long video viewing times from watching live videos or other content on your Facebook page — such as educational, how-to videos that are connected to your products or services.

4. Interest-based audiences

Even though we think we have a better success rate when starting with lookalike audiences for TOF campaigns, we also see success with interest-based audiences — especially when the product or brand has a precise market segmentation or personas. Aside from the product or niche interest groups, our usual process is to check the Audience Insights tab to find other things page visitors are interested in.

PRO TIP: When creating these audiences, our go-to process is to combine broad audiences with interests. We leave demographic attributes open and select 1-1 main interest only, which usually results in a massive 5M+ audience. 

Example: We target people in the USA with an interest in Crossfit, excluding our website visitors in the last 180 days.

Once you’ve analyzed the four categories above, choose 5-8 audiences you believe are the most valuable for a splash of fresh traffic. There are a couple of things you should keep in mind when creating a TOF campaign with Campaign Budget Optimization (CBO) and choosing audiences:

If the audience sizes between ad sets vary significantly — for instance, one audience has an audience size of 70k and another with 10m — then the algorithms will likely push traffic to the ad set with a bigger audience.

If you choose LLAs based on the same sale funnel events each time, you’ll more than likely find that they overlap. This overlap limits the algorithm’s ability to test fresh traffic and find you more conversions.

Let’s say you create a 1% LLA based on:

  • Add to Cart last 180 days
  • Initiate Checkout last 180 days 
  • Purchase last 180 days 

On paper, these all look like valuable lookalikes — and they are, to a point. However, when you want to give the best chance for a TOF Campaign Budget Optimization (CBO) campaign, you might want to mix up your audience and try new directions.

1% lookalikes are a solid go-to and will likely be more valuable than 2% or 5%. But we’d still recommend testing different percentages on the same source audience, as they might give you surprising results. 

Before you expand to broader percentages, we recommend that the first TOF campaigns only use a 1% audience.

Scaling your brand online is a challenging task, but you can make the path to growth easier by following strategies to affordably reach new customers. By implementing the processes we’ve outlined for finding and targeting cold traffic audiences on Facebook and Instagram, you’ll put yourself well ahead of your competitors.


Tired Of Unprofitable Campaigns?

Learn the method we’ve used to help profitably grow and scale e-commerce brands to 6-7 figures online with paid advertising. Save time, money, and countless hours testing by following our scaling method.


More To Explore

Ready To Scale Your Brand?

Let us take a look under the hood of your business and find 2-3 new opportunities for
growth, giving you the steps needed to scale and grow your business.​

Need help scaling your ad campaigns?