When writing one of my recent posts ‘Best Practice Guide: Facebook Advertising For Councils and Local Authorities‘, it highlighted just how important it is in Facebook advertising to split test your adverts. I have therefore decided to write a post on the importance of split testing, and how to best go about it.

One of the biggest reasons that brands fail in their Facebook advertising efforts is that they do not split test (Loomer, 2012). They create one or two ads, find that they do not succeed and label it a failure. That is why you must try out multiple variations in order to understand what resonates with your target audience. When you realise something that does not work, stop spending money on it, when you discover something that does, keep going with it.

What is split testing?

Split testing, known as A/B testing, allows you to test different elements of your advertising in order to determine which will help you to achieve your objectives (Loomer, 2012). It allows you to show different experiences to different users to measure the impact of those differences, and is thought of as a fundamental factor in obtaining value from advertising and growing your online business (Jenkins, 2014).

Split testing gives you access to data, which gives you a competitive advantage because the more data that you have on what your audience likes, the better your product/service can be (Jenkins, 2014). Many observers believe that receiving data through split testing helps organisations to ‘leapfrog’ and become the best in the class due to having the ability to make better decisions based on experience rather than intuition (McGuire, Mayika & Chui, 2012).

The question is what aspects of Facebook advertising can be split test?

The way in which you split test your adverts is dependent on the industry in which you operate and the type of products/services that you are selling. However, the following points, based on a guide written by Jon Loomer (2012) which can be found here, can be used universally and should work as a checklist to help you successfully split test your advertisements:

1) Copy

In Facebook standard adverts, you have 90 characters to persuade a user to complete a desired action (Facebook, 2015). This is a small character limit, meaning that you will need to test different variations in order to understand which type of messaging is going to be effective at persuading users to convert.

Standard advert example:

Source: Facebook (2015)

It is slightly different if choosing to advertise through sponsored posts, as the character limit for the body is 500 words; therefore you have an increased amount of words to convey your key message to your audience. Facebook does, however, advise not to fulfil the 500 word limit in order to keep the message short and catchy (Facebook, 2015).

Facebook sponsored post example:

Source: Facebook (2015)

Create adverts and switch up the copy through using different words, call to actions, tones and punctuation. Let’s say that you are a local authority or council seeking to recruit foster carers, but you are unsure on the type of messaging needed to encourage users to convert. Create three different advert variations, each with different content and a different call-to-action. For example your first advert may focus on the support and training which would be provided when becoming a foster carer with your organisation with the call to action ‘Find out more’, whereas the second may focus on making a difference to a local child’s life with the call to action ‘Register your interest’.

2) Imagery 

We already know that imagery is the key to advertising success on social media, as your advert is competing for attention against user friends and family (Facebook, 2015). Imagery in advertising is so important in fact, that Rossiter (1980) believes that it has the potential to change user attitude towards completing a particular action.

What we do not know, however, is which images are likely to persuade a user to complete a desired action. It is important to remember that all users are different (Close, 2012) so the preferred imagery will be different from person to person and demographic to demographic.

Recommendations for imagery are to use different faces in different scenarios, as well as text imagery, in order to distinguish the type of image which resonates with your target audience.

3) Landing Page

Advertisers spend obscene amounts of money to buy traffic, but the effort that is devoted to the landing pages to which the user is sent is often negligible (Ash, 2008). It is important to remember that all of your hard work in creating adverts comes down to a few precious moments when the user arrives on your landing page.

You have created strong adverts through effective content and imagery which are encouraging users to click on the advert; however, users are not completing the desired action on your page. This means that your bounce rate is high, which impacts negatively on your conversion rate (Zigmond et al, 2009), therefore ultimately affecting your overall goal.

It is difficult to answer why users are or are not responding to your landing page (Ash, 2008), but there are ways to determine the type of content that encourages your audience to convert. With the same advert; create multiple landing pages in order to determine which is the most successful at increasing conversions after the initial click is made.

 4) Demographics 

You may have the right content, with the right imagery, but your adverts could be targeted at the wrong audience. You have the option to select different imagery and copy that will appeal to a specific demographic, or try out a generic ad and split test it to different age and gender groups (Loomer, 2012).

By running your adverts to different audiences you will be able to understand the age and gender which correlates with your product/service offering, therefore decreasing the chances of aiming your adverts to the wrong audience which can cause advertising wastage (Greenyer, 2004).

5) Detailed Targeting vs No Targeting

Facebook offers a range of audience targeting options to choose from in order to help organisation’s meet their goals. These options are custom audiences, lookalike audiences, location, demographic, interests, behaviours, education and connections (Facebook, 2015). For local authorities and councils seeking to recruit foster carers, for example, this means that you are able to target people who display interests in pages relating to fostering and caring for young people, for example ‘Fostering Network’ or ‘UK Children’s Charities’. Or for organisations in the retail sector, you are able to target people who display interests in pages relating to fashion and beauty.

These multiple targeting options provided by Facebook increase the likeliness of reaching out to the right people (Marcella, 2009), however they mean that it is sometimes possible to overthink your audience (Loomer, 2012). Targeting an advert too specifically means that you may potentially cut out a segment which is likely to respond, so it is important that you test your adverts at both a specific and broad audience in order to ensure that you are not missing out on a particular segment.

 The Importance of Continuous Split Testing

Many organisations believe that split testing advertising only needs to be done once in order to create value; however it needs to be carried out on a continuous basis. This is because things change a lot.

Technology is continuously developing and adapting, which is driving the need for organisations to change the way that they market their products and services. As the population spend more time on their mobiles, tablets and laptops, organisations are realising that they need to connect with their audience through these devices (Benady, 2014). Investment in digital marketing is continuing to grow, with 77% of organisations saying that they intend to increase budgets in 2015 (Charlton, 2015).

This increase in digital marketing intensifies the competition for organisations seeking to attract the attention of users, as there are a large variety of options available for them to choose from. As technological advances increase, users are now far more empowered and in control, and their tastes and preferences are changing on a continuous basis (Forbes, 2014). Therefore, split testing your digital marketing efforts needs to be carried out on a continuous basis in order to stay ahead of competitors and meet the needs of your audience.

Concluding Points

Split testing adverts is an extremely beneficial way to understand how you can best reach your goals (Loomer, 2012), as it allows you test varying factors in order to understand both your audience type, and the type of content that resonates best with that audience. The data captured from split testing creates competitive advantage as you are able to continuously make your product/service better based on experience rather than intuition (McGuire, Mayika & Chui, 2012).

It is, however, important to remember that split testing needs to be carried out on a continuous basis in order to keep up with advances in technological systems and changes in user behaviour, tastes and preferences (Forbes, 2014).

To find out more about how you can best split test your Facebook adverts, watch the video below:

Source: YouTube (2013)

For further reading on the topic, please see the references below:

Ash, T. 2008, Start Thinking About LANDING PAGE OPTIMIZATION, Penton Media, Inc., Penton Business Media, Inc. and their subsidiaries, Rockville [accessed 09 April 2015]

Benady, D (2014) ‘How technology is changing marketing’ The Guardian, 29^th September 2014 [Online] <http://www.theguardian.com/media-network/media-network-blog/2014/sep/29/technology-changing-marketing-digital-media&gt; [accessed 09 April 2015]

Charlton, G (2015) ‘77% of businesses plan to increase digital marketing budgets this year’ Econsultancy, 26^th February 2015 [Online] <https://econsultancy.com/blog/66135-77-of-businesses-plan-to-increase-digital-marketing-budgets-this-year/&gt; [accessed 09 April 2015]

Close, A. (2012) Online consumer behavior: theory and research in social media, advertising, and e-tail, Routledge Academic, London [accessed 09 April 2015]

Facebook (2015) Ad Set Audiences [Online] <https://www.facebook.com/help/433385333434831/&gt; [accessed 09 April 2015]

Facebook (2015) Clicks to Website: Links [Online] <https://www.facebook.com/business/ads-guide/clicks-to-website/links/&gt; [accessed 13 April 2015]

Forbes (2015) Six trends that will shape consumer behaviour this year. Forbes. 02^nd April 2014 [online] <http://www.forbes.com/sites/onmarketing/2014/02/04/six-trends-that-will-shape-consumer-behavior-this-year/&gt; [accessed 09 April 2015]

Greenyer, A. (2004) ‘The impact of different media channels on consumers and the wastage of potential advertising opportunities through existing customer communications’, Journal of Financial Services Marketing, vol. 8, no. 3, pp. 279-290 [accessed 09 April 2015]

Jenkins, W (2014) ‘A/B testing and the benefits of an experimentation culture’ Harvard Business Review, 05^th February 2014 [Online] <https://hbr.org/2014/02/ab-testing-and-the-benefits-of-an-experimentation-culture/&gt; [accessed 09 April 2015]

Loomer, J (2012) ‘Facebook advertising: 6 ways to split test like a pro’ Jon Loomer, 07^th August 2012 [Online] <http://www.jonloomer.com/2012/08/07/facebook-advertising-split-testing-infographic/&gt; [accessed 09 April 2015]

Loomer, J (2012) How to split test Facebook ads [YouTube video] < https://www.youtube.com/watch?v=ZVTJqTQDB7w&gt; [posted June 06 2013 accessed 13 April 2015]

Marcella, V (2009) Critical Evaluation of Internet Advertising. Germany: Druck und Bindung.

McGuire, T., Manyika, J. & Chui, M. 2012, “WHY BIG DATA IS THE NEW COMPETITIVE ADVANTAGE”, Ivey Business Journal Online, , pp. N_A. [accessed 09 April 2015]

Rossiter, J.R. 1980, “Attitude Change through Visual Imagery in Advertising”, Journal of Advertising, vol. 9, no. 2, pp. 10-16 [accessed 09 April 2015]

Zigmond, D, Sundar, D, Interian, Y, Navemiouk, I (2009) ‘Measuring advertising quality based on audience retention’ Journal of Advertising Research [Online], Vol 49 (Issue 4) [accessed 09 April 2015]