[AUDIO VERSION: I also recorded an audio version of this blog post. Click below to listen. Let me know if this is something you find helpful!]
When you analyze your Facebook advertising reports, are you focusing on all of the right metrics? Many aren’t, even when their base metrics are correct.
A prime example of this is with website clicks. Recently I’ve been seeing that I can get a very cheap cost per website click when targeting Lookalikes and interests. In fact, it’s been far lower than the same costs when targeting my fans.
But how could this be? It could mean that either Facebook is doing an insane job assembling Lookalikes and interests. It could also mean that the quality of my fan base is not what I thought.
That’s why it’s important to dig beyond the website click to determine the quality of those actions.
[Want to master Facebook ad reports? It’s one of the featured topics in my 2015 FB Mastery Workshops! Go here to sign up.]
During the past few weeks, there were two main posts that were at the center of this experiment:
In each case, I created four different ad sets, one for a different audience:
- Fans: 25-49 in US, UK, Canada and Australia who speak English
- WCA 30: Website Visitors (past 30 days) 25-49 in US, UK, Canada and Australia who speak English
- Lookalike Audiences: Based on Fans and Website Visitors 25-49 in US, UK, Canada and Australia who speak English
- Interests: Mari Smith, Amy Porterfield, Facebook for Business or Social Media Examiner 25-49 in US, UK, Canada and Australia who speak English
Ages and countries are based on what I know about my current customers. I based my Lookalike Audiences on groups that I regularly target and who convert. The interests were determined using Audience Insights as those that are most common among my audience.
I also used URL tags so that I could track those who clicked the link from each ad to determine what they did while on my website. For example, the campaign name for the tag may be “quitlookalikes” for the ad reaching those people identified as a Lookalike Audience.
I also excluded people who would have already read the blog post using Website Custom Audiences to eliminate waste.
For each campaign, I created ads that promoted an existing post. I split up spend between each audience as evenly as possible.
Quit Facebook Post (Spend)
- Fans: $136.17
- WCA 30: $135.95
- Lookalikes: $136.28
- Interests: $136.21
No More Promotional Posts (Spend)
- Fans: $71.04
- WCA 30: $71.03
- Lookalikes: $71.04
- Interests: $71.04
For the remainder of this blog post, I will combine results for identical audiences. For example, total spend targeting fans is $207.21 ($136.17 + $71.04).
- Fans: $207.21
- WCA 30: $206.98
- Lookalikes: $207.32
- Interests: $207.25
Cost Per Website Click
The primary purpose of these campaigns was to drive traffic to my website. Here are the total number of website clicks and costs per website click for each audience:
- Fans: 749 Website Clicks ($.28 per Website Click)
- WCA 30: 1,416 Website Clicks ($.15 per Website Click)
- Lookalikes: 1,635 Website Clicks ($.13 per Website Click)
- Interests: 1,242 Website Clicks ($.17 per Website Click)
Since my primary objective was driving website traffic, it would appear on the surface that my most efficient audiences were Lookalikes and WCA 30. Even Interests were in the acceptable range.
On the other hand, it cost me $.28 per website click when targeting fans. This is a group that I target regularly. They are those most likely to engage with my post. Yet, it would appear that targeting them this time was a complete waste of money.
Keep in mind that my primary objective here was to drive website traffic to blog posts that were not positioned to convert. I did, however, have both a pop-up and top right widget promoting my ebook. So a percentage of those who visited these posts would undoubtedly convert.
When I created these campaigns, I also tracked pixels associated with this ebook so that I could see how many of those who clicked the ad to read my post ended up converting as well.
Following is the number of ebook conversions by audience:
- Fans: 22
- WCA 30: 19
- Lookalikes: 3
- Interests: 9
That’s crazy, right? Before we saw that fans resulted in fewer than half the website clicks of Lookalikes, yet they converted at a far higher rate.
Here’s a look at the percentage of website clicks that resulted in a conversion:
- Fans: 2.9%
- WCA 30: 1.3%
- Lookalikes: 0.2%
- Interests: 0.7%
Something to keep in mind here is that fans and recent website visitors are far more likely than interests and lookalikes to have previously subscribed to my ebook (more than 10,000 people have subscribed). As a result, a percentage of fans and recent website visitors couldn’t have subscribed again.
To summarize: It cost me far more to drive website traffic when targeting fans, but those people ended up being far more likely to convert. On the flip side, I could drive lookalikes to my website very inexpensively, but conversions were virtually non-existent.
Time on Site
Since I was using URL tagging on the links in my ads, I can break down the amount of time people spent on my site who clicked those links with the help of Google Analytics.
Full disclosure: A small percentage of those who clicked the links wouldn’t have viewed my ad. If someone shared my ad to Twitter or somewhere else, the link could have then been clicked by someone not in my target audience. Still, that percentage will be small, so this analysis remains valuable.
Here is a breakdown of average time on site by audience:
- Fans: 1:12
- WCA 30: 0:55
- Lookalikes: 0:26
- Interests: 0:33
Based on conversion results, this shouldn’t be a surprise. Fans spent more than twice as long on my site than Lookalikes and interests. My recent website visitors weren’t far behind, spending more than double the time of Lookalikes and 0:22 more than interests.
Dig Beyond the Click
This is an example of why you need to dig beyond the click when analyzing your results. I’d typically say that someone focusing on Cost Per Website Click when traffic is the objective is looking at the right metrics, but clearly that’s not always enough.
My instinct has always been to target those who are most closely connected to me first. That’s why I always target fans and WCA 30. I rarely go beyond that since I have less trust in the quality of the audiences when using interests and Lookalikes.
But if I didn’t know any better, I would have stopped my ads that targeted fans. In fact, I may have escalated budget on the ads targeting Lookalikes.
As you can see, that would have been a huge mistake. Interests and Lookalikes were giving me empty clicks while fans and recent website visitors provided quality visits.
In the end, the quality of actions you’re measuring is most important!
Have you broken down your results in a similar way? What are you seeing from these four different audiences?
Let me know in the comments below!