How to test and optimise your social media advertising campaigns

Facebook advertising

Facebook advertising should be tested and optimised just like any other channel

As every marketer will tell you. a huge aspect of running any channel is testing, learning and optimising. No campaign is the finished article and marketers should be in a constant state of learning.

Even if a particular channel is performing well, efforts should always be made to optimise. Why settle for good when you can have brilliant?

Social media advertising is definitely no exception and brands should be making every effort to constantly tinker and improve social media advertising activity.

In addition to on-going optimisation there are factors which can have an impact on the day to day running of a channel. Such external variables include time of year, time of day, industry developments and the activity of competitors.

A/B testing is the foundation of optimisation

A/B testing is the standard process for testing and optimising digital marketing channels and social media is no exception. Essentially it consists of making a duplicate version of an ad/landing page and running both simultaneously to see which performs better. The percentage split of traffic seeing each type can vary. If traffic volume is high enough a smaller ratio can be used (say 80:20). After a predetermined period of time the results are analysed and the ‘winning’ option is made the default option.

Best practice testing is to continuously A/B test channels and campaigns. Social media is no different and ad content should be constantly tinkered with and improved. Below are some suggestions for what should be tested on social:

  • Test a mix of creatives

When launching advertising campaigns on social media it is tempting to build one or two creatives which fulfil your need and then leave it at that. However, it is extremely difficult to predict what will work and what won’t. As with everything the best way to achieve and maintain success is to ensure creatives are A/B tested regularly.

Different people react to different things. Different images, colours and tag lines cause different feelings in different people and the best way to see what’s working, what isn’t and where any opportunities may lie is testing. Creating multiple creatives right at the beginning is sensible but it is also important to be able to tinker and change creatives during campaigns.

The range of different creatives should be quite wide and what is tested should vary. While small background changes may work it is advisable to A/B test larger changes so that you can investigate different opportunities and options without pigeonholing the campaign. The key here is to resist the temptation to think that what you have is the finished article. Keep your mind open and try different things – the A/B tests will tell you what works.

A word of warning however – avoid testing 1000s of creatives unless you have the budget to match. Testing only works if you have enough data to make it statistically relevant. The performance of an individual ad should not be determined based on handful of interactions and impressions.

  • Test devices

There are multiple ways to login to social media and advertising differs depending on the device people are using to br

owse. Screens are smaller (or larger), scrolling is different and content is consumed differently depending on device. Because of this, advertising campaigns on social media should be tested by device to see how different device users react to different types of ads, products and creatives.

It could be that your customers are over represented on different devices or that your creatives and customer journey is better suited to tablets.

You may also find different creative works better or worse depending on screen size so don’t forget to incorporate this into the testing.

  • Test call to action buttons 

As with creatives, call to action (CTA) buttons should be tested and changed constantly. In some cases the CTA on the advert may be self-explanatory, i.e. you wouldn’t use ‘shop now’ if the action is to have people sign up on site. Other options would include ‘learn more’ or ‘sign up’ buttons.

The simplest test is whether you have a CTA button at all. It may be the creative is so strong it doesn’t need the button to illicit a response or perhaps it provides added value to the users.

  • Test lookalike audience vs custom audience vs website audience

With the advent of audiences on Facebook, the capacity to focus on more specific customer types has increased in the past 12 months. It now includes many options for advertisers that want to reach as many relevant users as possible.

What this means however is that each audience created may react differently to advertising on Facebook.

For example a custom audience is usually made up of an advertiser’s current base of customers. So this audience would be best served with direct response ads vs. brand awareness.

However, a look-a-like audience may get a better response with brand advertising as while they look like your current customers, they may not be aware of your brand.

The same applies for website custom audiences which can be compared with retargeting. It could be that offers (for instance) will resonate with a customer that has come to your site but left for currently unknown reasons.

Don’t stay standing still

Marketing is a competitive environment – regardless of the industry. Without continuously optimising channels it’s possible to quickly fall behind competitors. Testing is the foundation of optimisation as it provides the learnings to make informed decisions. Best practice optimisation is to adopt an on-going schedule of tests and changes.

Even if a campaign seems to be working it’s always worth continuing to optimise – it can always be better.

Experian Marketing Services and the Alchemy Social platform makes social media advertising quicker, easier and more efficient.

If you need help uncovering your audience, Experian Marketing Services offers a variety of customer insight solutions.