How to Decrease your Cost Per Conversion by Doing Split Tests on Facebook

Recently I launched an ad and I made a mistake – I didn’t actually do a split test of the campaign before I launched. I created a cool video (which by the way – looking back was not the best option in hindsight as my goal was conversion and videos drive more awareness and engagement gaols) and I targeted a pretty good audience – being people that had visited my client’s website in the past.

The ad did well in getting about 50 cents per click (all though no conversions – however the button clicked was ‘Book Now’ so I can only assume that a click showed some intent).

However looking back I wanted to talk about how I setup a split test based on the reports and why Facebook is awesome when it comes to getting audience insights after a campaign is launched – unlike Google.

With Google – if you launch a campaign you can tell which people searched for which keywords – but at the end of a campaign you can only garner conversion results from specific keywords and optimise as so. So you could say:

“Well this keyword ‘x’ brought in this much conversion for our ad spend, but this keyword ‘y’ brought in this much conversions – so this keyword is better based on spend vs. results”.

However the big advantage Facebook has is giving you insights that you had never even considered.

For example in my latest campaign – after launching it I noticed a very clear advantage in the amount of engagement/clicks between women and men (women tended to click the ad at a much higher rate).

This type of demographic reporting is simply not possible with Google – and all though Google is great at getting customers at the moment they are searching for a solution – the insight you get is much lower. 

The great thing about Facebook is how easily they allow you to split test campaigns based on different variants. In fact the tools are built right into the system with one click solutions – you just click an ad set, ad group or ad campaign – and click on ‘Duplicate’. You can then run the variants to see which performs better.

What’s great is that once you find various target groups you can then target your ads based on those groups – so while my next campaign will only have the gender variant (one campaign is targeting all genders while the other is targeting females only) – later I could tailor an ad to only women – since I am promoting a guest house at a spiritual retreat – to have photos of a woman doing yoga or something along those lines.

To do a split test simply click on ‘Duplicate’ > ‘Quick Duplicate’ at either the ‘Campaign’, ‘Ad Sets’ or ‘Ads’ level. Creating a variant then involves simply changing some minor point on either of the rows in the categories above. And voila – you’re off and running.

Running a Facebook campaign without a split test is not a good idea – and there has NEVER been a time when – prior to planning a campaign people didn’t have a disagreement on the best way to setup ad creative or targeting – now the guesswork is out of it – it’s easy to setup and test the results.

This means that even if a campaign is not as successful as you would have liked – you can almost treat it like a market research excercise to find out what doesn’t work and what demographics are more engaged with what creative.

If you are looking to launch a split test and get engagement with your product/service leave your details below and I can launch it immediately in my business manager and let you know about your customers. 

Let’s do this!