A simple targeting strategy that brought 58k leads at less than $5 – Facebook Ad Campaign

Last night, I stumbled upon our old Facebook ad account. I kept looking at for a while with a surprise that we seem to have done a better job of managing that client account. 

It’s a property industry. If the user converts, his house price would be minimum of USD100K. We are not allowed to access that level of data due to confidentiality. However seeing the client wants to continue spending and their verbal acknowledgment, we concluded it’s working and continue to promote on Facebook.

A lot of experimentation, we are able to make it. We didn’t do anything big. A simple strategy that has given the steady flow of leads with so much engagement on social media. The engagement here means the comments and questions on private messages. 

Just the leads ads campaign

This is what I have done. 

Our situation is we don’t have any existing user base or anything. However the brand has good trust in offline world. 

Profiling Analysis

This is the first step I do for every client. But this time didn’t do. Due to the vast variety of property types they had, I didn’t do so much on profiling or any granular segmentation of who we want to target. 

Before I ran the campaign1, the client has been running similar campaigns by themselves and also more spending on Link Clicks (it doesn’t work from what I found). They have tested a few lead ads campaign as well.

Campaign1: Lead Ads (Interest)

The simple segment we started is 35-65 (hypothesis is should have buying power). And have target interest such as Yatching, Private property, Yacht club, Luxury real estate, Creative real estate investing, Golf course, Investment, Real estate investing, Golf, Real estate, Golf club or Real property

This brought as 2-3K leads easily at the same price I mentioned above for one property project.

Campaign2: Leads Ads (Interest + Lookalike)

Then another property campaign. The same segment as campaign1 but with some additional interest like wealth management, commercial property, etc 

This brought like another 4K leads. 

Nothing big in terms of creatives. Just simple Link leads Ads with a sky view of the whole property. Good amount of share. 

From the first campaign, sort of got the idea of how much is the cost per lead and related metric. Then I duplicated the same audience + creating lookalike as well but with 

  • Manual Bid of 20 per lead target, 
  • Manual bid of 30 per lead target. 
  • Manual big of 25

Result of this bid testing came clear after campaign 3.  

Campaign3: Lookalike focused

From the initial 2 campaigns, I already got a significant seed audience. We used that to create lookalikes. 

The campaign targeting was not 1% of lookalike but 10% of a lookalike. Here I want to highlight of confusion about the size of the audience where each facebook marketer has. 

Small Size: The misconception that the smaller the size, so it’s easy to decide whether a certain audience is working or not. 

Bigger Audience Size: The misconception When It becomes too broad which is not easy to control. However, the bigger size audience(not more than a million) seems to work better than the smaller. Am not completely sure how facebook optimizes but my hypothesis is smaller the audience, higher the competition and higher the bid. So when it’s broad, facebook machine at a time can find low competition audience based on bids. 

So when I find the lookalikes were working, I tried the bidding experiment as shared above. 

“Manual Bid of 20” ad set start to work better than the other. I continued with it. 

Bidding Matters if you have already tested the audiences

Closely watch the Mbid and their Cost per Lead (CPL)
Look alike Mbid21 (it’s CPL is RM20), whereas Mbid30 CPL is 27
Here the result is varies for Mbid for Broad vs Lookalike. For lookalike it’s clear that bid proportional(RM20) to the cost per lead(RM17)
Here you can see all the ad set bid is 35, then the cost per lead 30+

May be I could have done it better than as well. However, the client is more than happy about this. Other things I did

  • I didn’t frequently change anything (not within 24hours, I allowed to get leads first to gauge correctly)
  • I tested with few images for some property(like carousel ads) when we see link ads were not working) (This is property project level and not overall)
  • As I seen the lookalike were working better, than competitor interest like “Property company1, Property company2, etc”. I stuck after a while with one broad audience and another lookalike with their split
  • We kept adding the new leads into existing custom audience to make sure we didn’t target existing users.
  • Also when the seed size is improving, I created new lookalike as well. Not frequently but once in a quarter and twice.
  • Also for new campaign creation, we duplicated existing ad set which is working without doing a drastic change in the targeting