In a recent Facebook study done in conjunction with Ipsos Market Research, they found that “creative quality determines 75% of impact as measured by brand and ad recall”.
Delivering the right creative concepts for audiences is the most important factor in driving performance (clicks, conversions, traffic). In fact, without the right creative, other campaign settings such as audience targeting and ad placements won’t help you drive more conversions and engagement.
But wait a minute – with ad platforms offering sophisticated optimization tools for A/B testing and live experimentation, aren’t we already pre-testing creative?
The simple answer is no. Pre-testing creative involves testing creative concepts against your target audience before deploying for live testing.
Standard A/B tests and live experimentation are traditional ways to identify winning creatives. But, there are 3 main problems with this approach, and they share common issues in the inability to comprehensively and objectively benefit from harnessing all the underlying historical performance audience data.
1. Wrong creative concepts in the first place.
As the ad platforms grow in capability and automation, more of us leave the heavy lifting to the platform. The problem is that this hinges on you having the right creative to begin with. Testing out all creative concepts before deploying live will save a lot of time and budget.
Utilizing Facebook and Google optimization insights are critical to increasing impressions to audiences, but do not identify whether your ad creative is right for your target audience. You only find this out after the fact, after you’ve spent time testing and wasting ad spend.
It can also explain mistargeting. Most brands have experienced this, settings and targeting that bring in a lot of traffic, but often traffic does not boost the bottom of funnel impact. You can achieve a highly optimized grade from the platforms but still underperform or mistarget completely.
2. A gap between the brand strategies and what generates performance
Without a bottom up data analysis, we go into a new campaign partially blind, and are more than likely to underperform on the potential opportunity. Filling in these blanks is challenging, all too often a process of costly trial and error, and open to human bias because we lack objective and complete analysis.
Until recently, marketers could not easily comprehensively analyze their audience data for creative pre-testing. It’s not hard to understand why either. Anyone who has spent time on a Google or Facebook ad metrics dashboard, or any ad platform’s metrics dashboard, knows the challenge of using the vast underlying data — it’s a bit of a black box.
With an audience of thousands, let alone millions, all interacting on a platform like Google or Facebook, the volume of data captured is massive. No human data scientist, let alone a pressed-for-time marketer, can query each and every data point to discover all the key relevant relationships and concepts for a winning ad creative.
As a result we simply miss key insights on copy and images that have a powerful impact on audiences, and can define specifically what drives demand.
3. Human bias and subjectivity can limit creative analysis and competitive differentiation
In the past, pre-testing creative meant live experimentation and A/B testing, or even focus groups. These can be time consuming and costly, and more critically, severely limited by the human assumptions as to what to test and how to interpret the results.
With a traditional data analytics approach, we rely on generating sensible assumptions about our audiences, then test and prove our assumptions. This approach means recognizing that these assumptions could be flawed or incomplete.
Too often campaigns coalesce on similar concepts, or rehash previous creative, reducing the risk. This impacts competitive differentiation, generally our assumptions are likely to be based on broadly shared experiences, meaning the marketers running competing campaigns are also applying the same approach.
We simply cannot manually generate assumptions that test every data point and combination of data points to discover critical audience-creative relationships. There is not enough time or money to have enough data scientists on hand 24/7 to crunch all that data.
A Job for AI Machine Learning
With advanced AI machine learning capabilities, including image analysis, brands now have incredibly powerful tools to help boost a campaign’s likelihood of success, and critically, avoid wasting time and funds on underperforming ads.
With an AI l-enhanced solution, we have the opportunity to pre-test multiple creative concepts on our audience data and get instant feedback. This provides a more comprehensive analysis, as the AI can derive all data from the creative – every element in the image and word in the copy – and crunch vast amounts of data in no time, to provide pre-test predictions and insights.
At Junction AI, our testing has shown that audience models using 1st party data are critical to teasing out the creative insights that are relevant to a brand and that differentiate from competitors. The depth, breadth, relevancy, and critically, objectivity backing AI generated insights cannot be matched by traditional testing approaches.
How does this power campaign performance?
Pre-testing creative with an AI offers the opportunity to test all data and relationships comprehensively, so we can discover the concepts from the ad creative that represent what drives engagement. This is like generating a tactical definition of a brand’s marketing strategy.
With the results of an AI-enhanced pre-test, a marketer can fill in the blanks between the concepts that an audience is looking for and want to engage on and the higher-level marketing strategic messages.
We’ve also taken this to the next level using our Advertising Intelligence™ solution to pre-test competitor ad creative. Testing on competitor creative can uncover insights on a competitor’s ad that are most attractive to your audience too.
Plus, a competitor test is a great way to enhance the internal review, to dispel any myths, assumptions or disagreements on what we subjectively interpret as competitor strengths and weaknesses.
All too often, we find a disconnect in what the brand sees as their core value proposition versus what the creative means to the audience. This disconnect costs not just the time and money spent on developing ad creative and the campaign, but more critically the lost opportunity – an audience that did not convert or who went to the competition as they “connected” with their creative instead.
Equipping marketers with the ability to rapidly pre-test creative on their own audience with an AI means marketers are better informed on what concepts in their creative drive results.
With this information, and data objectivity, marketers can successfully unpack effective creative strategies with the tactical insights that deliver engagement. Pre-testing may even deliver critical feedback back up to improve overall brand and messaging strategy with what actually generates demand and why!