12th February 2023

How to test your creative

Our resident behavioural science expert, Aaron Wells, spoke on The Behaviour Change Marketing Messaging Masterclass about testing creative concepts before they go into the wild. The presentation covered both pre-testing and A/B testing.

Pre-testing creative concepts before launching a campaign is divisive within the creative community. On one hand, it makes sense to find out how appealing a concept is amongst a target audience, however often people don’t know what motivates them implicitly and therefore may give feedback upon reflection.

However, there are certain situations where we’d recommend pre-testing concepts:

  1. Sensitive subjects – e.g. sexual health, vaccinations, etc.
  2. Hard-to-reach audience – e.g. minority groups, younger people, etc.
  3. Different stakeholder opinions – if there’s no clear winner on the concepts presented

A/B testing on paid social platforms such as Meta is a no-brainer because we look at actual behaviour rather than a prediction of future behaviour.

Read on to find out more about how we test creative! The presentation covered:

  1. How insight informs the brief
  2. What makes a great creative idea
  3. Creative testing principles
Joanna Kosinska 1 Cmofspfso Unsplash

The creative brief
How research & insight informs the brief

Aaron spoke about an important part of the creative process – the brief. There are different types of briefs, however the two main ones are the Client Brief and the Creative Brief, each one having some differences, mainly who’s the brief intended for (the reader).

The client brief is obviously written by the client and is generally received by the agency Account Director and Strategist/Planner, and should cover:

  • Define the need for advertising
  • The organisational problem
  • Objectives
  • Target audience
  • Budget
  • Align of evaluation

The creative brief is usually written by the agency Strategist/Planner and is received by a creative team (usually a Creative Director and a Copywriter), and should cover:

  • Organisational goal
  • Target audience
  • Consumer problem 
  • Insight
  • Single-minded proposition
  • The action (think/feel/do)


Aaron talked about how important insight can be to creating emotive campaigns. Insights unlock the problem and lead you to a strategy or solution.

An observation is the 'what' and 'insight' is the why.  An insight focuses on why people are doing what they do.

A great example is the This Girl Can campaign:

This Girl Can Strategy

What makes a great idea?

Creativity is both an art and a science and can often be subjective. However, when judging creative work, we look at the following:

Is it Distinctive? Being outstanding at standing out and demanding attention.

Does it move me emotionally? The most effective advertising is right-brain (emotive) thinking.

Can I attribute it? Helping to build the salience of the brand (the extent to which the brand has built links to attributes within someone’s memory).

Jess Bailey L3n9q27zulw Unsplash

Creative testing – pre-campaign

After concepts are presented, often clients have an opinion on which one they prefer, however often they are not the intended target audience. In this case, it’s useful to put them into a testing environment. 

So, in some situations, we’ll pre-test creative quantitatively through a research platform to find out things like: meaning, likability, motivation and memorability.

We would then work up the concepts to a slightly more finished state and choose the audience we want to test against – age, location, and demographics. Then, we’d write the questions, make the survey live and wait for the results to come back. Our partnership with the University of Birmingham’s School of Linguistics means we get an expert Researcher to evaluate responses and provide recommendations.

Here’s an example process we went through when testing concepts for our client, Umbrella:

Creative Testing Plan
Pre Testing Creative Umbrella 2
Pre Testing Creative Umbrella

A/B testing on Meta

A/B testing lets you compare two versions of an ad strategy by changing variables such as ad images, ad text, audience or placement. Each version is shown to a segment of an audience, ensuring that nobody sees both, then determining which version performs best.

This is where you can test creative in a live environment, measuring how people engaged with the ad by clicks, for example. Aaron suggested having a checklist before you start testing on paid social advertising platforms like Meta:

  • Time and resources – do you and your team have sufficient time to carry out the testing planning and implementation? 
  • Creative team – do you have access to a creative team e.g. a Designer, Creative Director, Copywriter at an agency or in-house?
  • Research & insight – have you carried out your research and developed an insight
  • Methodology and hypothesis – from your research, do you have a hypothesis and a method for testing?
  • Select variables – what would you like to test (images, text, headlines, CTAs, etc.)
  • Review behavioural insights – have you considered reviewing relevant behavioural insights?
  • Paid media specialist – do you have access to a media buyer who knows their way around the platform?
  • Decide media budget/timings – create a timing plan and media budget

Using the below structure, you can think about what element you’d like to test, starting with adsets and ending with ads.

Meta Testing

One of the best examples of a brand that used A/B testing was KFC and Ogilvy in Australia. They wanted to find interesting ways of the $1 chips promotion using various behavioural insights. But they couldn’t change the product or the price, they could only change how people saw it (the frame).

In order to bring the right ingredients into the mix, they looked at the literature and found 18 different principles from psychology most relevant to the perception of value within fast-moving consumer goods.

They came up with 90 different ways to say $1 chips. The top 5 were tested against a control ad on Facebook using a measure of clicks over likes.

The top five were tested:

KFC Creative Testing

Reciprocity and anchoring significantly outperformed the control whereas social norming performed significantly below.

A subsequent radio and visual campaign used only anchoring and resulted in a 56% increase in total chips sales compared to the previous year and KFC saw an 84% uplift in 4 chips transactions.

Key takeaways:


The rate of unique clicks on sponsored social media posts was used as a proxy to access message potency. Consistent imagery was followed by five different articulations of the same message. 


KFC tested each of these five conditions plus a control one over the course of a week to see which one got the greatest response rate measured by unique clicks – they observed what engaged people and stopped them in their tracks to think about the offer.


A subsequent radio and OOH campaign resulted in a 56% increase in total french fry sales compared to the previous year. But it gets even more interesting when you look at where some of the sales were generated, the company saw an 84% uplift in 4 french fry transactions.

If you need some advice on creative testing, we’re more than happy to help. Drop us an email [email protected]

Sign up for our latest insights, ideas and inspiration