Ever since Facebook first launched its advertising platform, being able to target users based on demographics, interests, etc. was one of its key strengths over Google. Over time through politics, privacy issues, and countless other items, Facebook has pulled many of those interesting targeting categories away.
With the prioritization of campaign-level budgets, we’re able to now let Facebook better spend in areas that perform. That allows us to run some interesting comparisons between these highly refined audiences of the past and broader lookalikes.
Facebook is Smarter than Many of Us Give It Credit For:
As marketers, we should have a pretty good idea of who are target market is and where the majority of our customers come from. We can use that historical data to create some really refined audiences based on what people like, where they live, etc.
In the case of this example, we’re dealing with a brick and motor location. Here is a bit of background to help you understand what we’re looking at:
- Targeted zip codes are compiled from the highest performing areas around the store.
- Lookalike audiences are based off everything from website visitors to people who completed a lead form.
- We have mileage broken out to allow Facebook more leeway in allocating budget.
Here we can see a lookalike audience of people furthest from the store is actually driving the highest lead volume at a CPL barely above the strongest performer. Our carefully crafted “Targeted Zip Codes” audience has the lowest CVR and the highest CPL. If you originally started with only that audience, you may have been sorely disappointed with the results!
One of the biggest surprises of this test was the mileage radii. I would have never expected people further away, around a major city, would actually perform comparably to that closer population.
Bigger Can Be Better:
One trend I’ve been noticing over the last year is that I’m
achieving better performance from being less targeted than what used to be the
norm. Using a lookalike and letting Facebook do its thing is yielding better results
than slaving away to find the most perfect targeting. What this tells me is
that Facebook is really good at finding people who are more likely to
accomplish your ultimate objective.
By testing out a number of these audiences between
demographics, cascading negatives (1%, 2%, 3%, etc.), all in the same ad set,
you’ll be able to start getting a read on what works and those results may
Play to Facebook’s Strengths. Not Your Analytical Skills.
Rather than doing the front-end work to guide Facebook’s algorithm, we should shift focus to improving the results it can achieve. That comes mostly in the form of creative and other changes like account structure.
AI is becoming a large influence in our industry. As a result, being able to harness its potential across many platforms is becoming a larger portion of our jobs. At the end of the day, crunching tons of numbers from historical data may not actually prove to be more efficient than just creating a good lookalike audience.
If you’re continuing to use the same detailed targeting options on Facebook, I urge you to try something new and open things up. Let Facebook use its vast number of data points to find your most likely customers because it has access to data we never will!
This isn’t going to be the case with every account but it’s something to revisit if you’re looking to revisit your Facebook strategy.
If you’re looking for additional Facebook strategies to drive your 2020 strategy, check out our Guide to Facebook Advertising: Intermediate Edition.