My first Facebook Ads Experiment (Day 3)

Introduction

If you've been following my posts over the last two days, then you'll know what this is about. :)

For those that haven't, on Sunday I created an experimental Facebook ad by using the Boost Post feature on my most recent submission to my brand's Facebook page. It was basically a link to my latest comparison review of different courses within my niche and I was hoping to see some traffic and/or conversions as a result.

Setting the budget at £1 per day resulted in a total expected spend of £3.

I set the target countries to the UK and US (as I knew they were the main customers of my target product) and I chose three demographic categories to suit the target product.

Performance

As you can see from the banner image to this post, in total 333 people were reached.

Of those 333 people, 12 people engaged with the post in the advert and 10 people actually clicked on the advert. Two of the people that engaged with the advert liked the post on my Facebook page. The other ten presumably were routed to the post on my website.

Now here's where things get interesting.

According to the Google Analytics for the period of the campaign, I had 2 returning users. Not new users. Beween them, they had six sessions. This does not, in any way, reflect the 10 clicks on the advert. Of course, I don't know if two people that were already visitors to my site clicked on the advert but I think it's unlikely.

The bottom line is, Google Analytics doesn't seem to show any of the actual visits from my ad clicks.

Adblockers can screw up your data collection

So, where are my clicks? Did Facebook lie? Well, this is where an important fact about analytics comes into play. Adblockers, block Google Analytics requests. This means that when your website tries to tell Google about a website visitor, if the visitor is using an ad-blocker, Google will never know about them.

There are ways around it (which I have not implemented yet) but they could fall foul of the new GDPR rules the EU has implemented. At this stage, I'm not sure. Anyway, if you're interested in finding out more, there are pages that explain how to do it.

Don't believe Facebook's predicted reach

One area in which I feel Facebook misled me is the number of people that would be reached.

When I set up the ad campaign, Facebook estimated that my ad would be shown to between 180 and 940 people per day.

The 333 people I actually reached over the entire three days, was well below the lower end of the estimate. On average, I reached 111 people per day, not the minimum value of 180 as estimated. An estimate is of course a rough quote of expected performance but in this case, I think Facebook grossly underperformed.

Targeted demographic group

The Facebook campaign has proved useful in one aspect - it's shown me the sort of people that my niche appeals to as defined by my ad parameters. In the people reached graph below there is a clear indication that my advert was more appealing to women over 40. Only 35% of the people reached were under the age of 45.

That's interesting because it suggests that for a more successful campaign I could do two things:

• Tailor the post to appeal more to women over forty (I'd need to figure out how)
• Amend my ad's parameters to more accurately represent my targeted niche.
  

Summary

I've found this experiment to be interesting at the very least, if somewhat disappointing. Here's a round-up of my findings:

• I cannot confirm that I had any actual visits to my websites attributable to the Facebook ad
• Facebook's estimated reach is a meaningless statistic. Take it with a pinch of salt.
  
• Adblockers can seriously distort your actual website usage
• I did gain two likes to the Facebook post as a result of the ad campaign
• I was able to confirm that my niche's target demographic are women over forty years of age

I've opted to extend the ad experiment until the 27th May to see if the observations on demographic representation are consistent and to see if days such as Friday and Saturday produce better results, when people feel that 'end of the week' euphoria.

I will be producing a final summary post here when the experiment concludes on the 28th May.