Source: http://megaphonesociety.wordpress.com/

How To A/B Test Your Facebook Posts Without Spending A Penny

Find what message performs best on Facebook before wasting money on promoted posts

You’re probably already aware of the performance improvement you can achieve by A/B testing your Adcopy, or the messaging on your landing pages. You may even be already split-testing your copy on Facebook using the ‘dark post’ feature. However you probably didn’t know it was possible, using a combination of Bit.ly and the ‘hide’ button, to perform a quick and dirty test to find the right messaging before spending a penny on advertising.

Note: In order for this to work best you usually need around 20k fans

You can use this free significance calculator to ensure statistical significance

Step 1 — Track your URLs with Bit.ly

When you create a post on Facebook that has the goal of driving traffic to your website, you can use Bit.ly* to track clicks. To track two different links to the same URL you need to append something to the end of the URL to differentiate them. You can do this with utm tracking parameters (giving you more insight later in Google Analytics) but you can keep it simple by just using a dummy parameter. For example:

http://www.example.com/landingpage?r=ab1
http://www.example.com/landingpage?r=ab2

Bit.ly will see this as two different URLs and allow you to track seperately. (see the Bit.ly site for more information on how to do this).

Step 2 — Prepare your post as normal

Once you have your two Bit.ly links you should write the copy and pick an image for your post as normal. There are a tonne of helpful resources online to help you write posts that perform on Facebook, but here are a few basics to remember:

  1. keep the copy short and punchy
  2. upload a stunning picture
  3. have a clear call-to-action
  4. don’t give away the whole story in the copy
  5. don’t be afraid to ask for engagement

The following is a good example of a post that worked well for me:

This post has it all: short, punchy copy that doesn’t reveal too much, a clear link to click on, a prompt for engagement and (most importantly) a stunning picture!
Over 9,700 clicks!
Fun Fact: if you enter a Bit.ly link into your browser with a plus sign after you can see how many clicks it got: https://bitly.com/ThaiDeal+

Step 3 — Create the post you want to test

In Step 2 we prepared our post as normal, sourcing a stunning image and writing some great copy. Now we need to figure out what element we want to test. My recommendation would be to keep it simple and change just one element — this will give you a better understanding of what works so you can replicate it in future. Here are some things I’ve found to make a big difference in the past (in order of importance):

  1. The image
  2. The first line of copy
  3. The prompt for engagement
  4. Short vs long copy
  5. The link (bit.ly/TestThis)

Step 4 — Set up and run the test

Paste the post you think will lose into Facebook and let it run for 15 minutes. When time is up, count the number of clicks on the Bit.ly link (by adding a plus sign to the end and pasting it in your browser).

From the admin area, hide that Facebook post on the wall. Then paste the other post (the one you think will win), making sure you’re using a different Bit.ly link. Wait for 15 minutes then record the number of clicks as you did with the previous post.

If the first post (the one you thought would lose) got more clicks, hide the second post and post the first one again.

If the second post was the winner, you win! Consider promoting it to a wider audience if it’s hitting your performance targets.

Note: you can just delete the old post, but hiding it will keep all the interactions, so it’s better just to leave the hidden post there.

I’ve described how to do this on Facebook, but you can imagine a similar tactic would work on Twitter and other social media channels.

Credit: I picked up this technique from Upworthy’s Adam Mordecai in the comment section for this (amazing) slideshare presentation.

Contact me on Twitter with any questions / feedback @2michaeltaylor
One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.