44.83% increase in ROI with better lookalike audiences [Case Study]

If you’ve been reading my stuff for a while, you’ll know that a big tip I give out is to use your best customers/users/clients for your lookalike ‘seed’ audiences, rather than a large, general list.

For example, an eCommerce store could do this by creating a list of all of their customers that had high average order values, compared to all of their previous customers in general.

Today, I wanted to share a case study of that exact example.

The client in this case study is an eCommerce store that does around $15,000-20,000 a month in revenue. Initially, I had built a lookalike audience of all their previous customers in the past 6 months. In the screenshot below, this is the ad set on the bottom.

I then created a lookalike audience of their high AOV customers. As you can see, this resulted in a 44.83% increase in ROI and revenue.

If you can’t see the screenshot (I wanted to include all the data, sorry!), the original ad set stats were:

• Spend: $258.85
• ROAS: 2.90
• Revenue: $751.70
• Cost per Landing Page View: $0.70

Compare this to the high AOV ad set, which had:

• Spend: $258.71
• ROAS: 4.20 (44.83% increase)
• Revenue: $1,087.51 (44.83% increase)
• Cost per Landing Page View: $0.59 (15.71% decrease)

Clearly, the data you feed into Facebook has a massive result on the type of results you get. Since I’m ‘telling’ Facebook to find me more people like the people that have high average order values, Facebook has a better idea of what sorts of customers I’m actually looking for. And the results speak for themselves.

As I mentioned, in the case of an eCommerce store, segmenting down to your best customers is fairly simple. Simply target customers that have high average order values (AOVs). This client sells products in the $6-$8 range, so I created a .csv file of all their customers in the last 6 months who had AOVs of $20 or more.

Other ways you could do this:

• Repeat customers (purchased 2+ times in the past 6 months)
• Most recent customers (purchase in the last 30-45 days)
• Remove bargain shoppers (exclude everyone who used a coupon in the last 6 months)
• Customers in your top-performing country (eg. only US customers, for companies that ship worldwide)

Hopefully this gets the gears turning in your head. If this isn’t enough to convince you to try segmenting down to your best customers, I don’t know what will be.