Data is rewriting creative - and scale is the new battlefield

Creative production is now the key driver of advertising performance. You need to launch a high volume of quality creatives, identify the top performers, and optimise them to maximise results – all within a matter of weeks. No brand can do this consistently, and at scale, with creative intuition alone. You need to apply data to every stage of creative production and automate analysis to inform impactful decisions without delay.

Creative is the new targeting

AI-powered campaigns like Google’s Performance Max (PMax) and Meta’s Advantage+  automatically test creative assets to find the most effective combinations for each target audience. Instead of manually creating (and managing) campaigns for every target audience, you can prioritise creative production.

This allows advertisers to experiment more, discover what works, and refine results. With a data-driven testing system, the end result is a higher volume of winning creatives. Analysing over £140M in media campaigns across paid social and search, we consistently see the same patterns:

• A strong correlation between ad spend and the volume of new creatives launched

• New creatives deliver peak performance for only a few weeks

• Consistent creative testing drives more efficient, scalable growth

We’re not the only ones seeing this, either. Meta research finds that advertisers running at least 15 experiments per year see results improve by up to 30%. And, this increases to 45% after two years of consistent testing.

So, how do you produce a high volume of winning creatives during those first few weeks of peak performance? You need a data-driven creative system that quickly produces and identifies winning creatives.

Now, let’s look at how we apply data to each stage of this creative process: research and ideation, testing, and analysis.

Research and ideation

The goal of research and ideation is to uncover insights about your audience, market, and competitors to inform creative decisions. There are three core components to this:

1. Consumer sentiment & trends

2. Audience insights

3. Market & competitive landscape

These insights identify the pain points, market trends, competitor weaknesses, etc., that you use to deliver creative assets that resonate with target audiences.

1. Consumer sentiment & trends

Consumer sentiment analysis reveals how target audiences feel about your brand. While trend analysis reveals how sentiment changes in response to market trends and brand actions – eg: new product releases, PR campaigns, and viral posts.

You'll get the most in-depth insights by combining multiple data sources.

• Consumer trend reports: Provide a general overview of consumer trends that you can test against your own data to see where audience segments match or deviate.

• Customer feedback: Incorporate customer feedback from online reviews, surveys, support tickets, social media interactions, and all other available channels.

• Social listening tools: Monitor online chatter about your brand, competitors, and relevant keywords to understand general sentiment and identify any changes.

With the latest AI tools, you can collect and analyse this data faster than ever. You can use AI text analysis to extract key insights from customer reviews, phone call transcripts, social media posts, and other sources. You can also use tools like ChatGPT to summarise consumer reports, market trends, identify audience pain points, etc.

Now, you can get the insights you need to launch new creatives in hours, not days (or weeks). And, this is scalable, as you continue analysis, and launch each new round of creatives.

2. Audience insights

Audience insights make every creative you produce more relevant and compelling to its intended audience. Before you spend big on consumer research and audience analysis software, make sure you’re getting the most out of your existing data from customers and advertising platforms:

• Customer data: Analyse the behavioural data of existing customers. Dig deep into cohorts and shared behaviours to better understand different customer groups and how closely they align with trends.

• Audience insight tools: Pull data from relevant audience insights tools like Facebook Audience insights, Google Insights Finder, TikTok Insights, etc.

• Campaign data: Analyse existing campaign data to see how audience insights translate to campaign performance.

Again, you can use AI tools to speed up the analysis of multiple data sources to inform creative ideas. For example, you can ask your preferred tool to generate a bar chart segmenting customers based on their primary purchase motivation.

AI can analyse and visualise actionable data instantly. In the example above, this gives us seven ultra-relevant concepts to explore for creative ideation. Most importantly, you’re starting with creative motivation backed by customer data and the key motivators that compel them to buy.

3. Market & competitor landscape

Market and competitor analysis is a constant source of creative inspiration, especially as new trends and disruptors emerge. Again, you’ll need to combine data sources to get in-depth insights quickly:

• Market analysis: Get baseline insights for general market trends, so you can compare brand and competitor insights for similarities and differences.

• Social listening: This time, to track competitors – what they post, how audiences engage, their top-performing creatives, etc.

• Ad libraries: Most of the big advertising platforms provide a searchable library of active ads that you can use to search and view competitor creatives – eg: Meta Ad Library and TikTok Ads Library.

• Competitor analysis: Track competitors, their creatives, and performance over time.

To automate analysis, you use tools like ChatGPT to generate lists and summaries of competitor brands. But, you really need to track these brands over time to understand what they’re doing, what’s working, and how they react to market changes.

There are plenty of tools that can do this. We currently use MagicBrief, which tracks competitors and analyses their creatives to provide actionable insights for creative production.

With this combination of tools and analysis, you can launch a higher volume of winning campaigns faster. And, this allows you to dedicate more resources to testing and analysis during those first few weeks where it matters most for new creatives.

Testing and analysis

The goal of creative testing is to discover which ideas, messages, and visuals best engage your audience and drive the highest return efficiently. This incorporates two key phases:

1. New concepts (large swing): Brand new creative ideas, inspired by insights, that are fundamentally different from previous creatives.

2. Variations (iteration): Smaller, incremental tweaks to key assets in winning creatives: headlines, hooks, CTAs, imagery, colours, etc.

Producing a high volume of new concepts (large swings) unlocks hidden opportunities, pushes reach, and unlocks new angles that resonate with audiences.

Then, producing variations (iterations) for the key assets in winning creatives improves and sustains performance during that early peak cycle for new campaigns. To reach the variation level of testing, you want to identify winning creatives as quickly as possible. That period of peak performance is short, so you need to know which metrics correlate with meaningful engagement:

• Thumb spot rate (hook rate): Does the creative stop audience scrolling?

• Hold rate: Does the creative hold audiences’ attention after the initial hook?

• Click through rate: Does the creative make audiences want to click through?

• CPA, ROAS, CvR: Does the creative convert?

A winning creative will perform across all four of these key metrics. Just keep in mind that it takes a little longer to get reliable data for CTRs, and especially for CPA, ROAS, and CvR. So, make sure your bidding strategy allows you to collect enough data for both levels of testing during the first few weeks of new creatives.

Once your reporting system identifies a winning creative, determine why it works. What makes that hook capture and hold attention, and what characteristics make the key assets (headings, CTAs, etc.) so effective? Feed these insights into creative production. Test variations using the same hook, experiment with KSPs, and tweak the messaging to squeeze out every bit of performance.

Applying data to creative at scale

Applying data to every stage of creative, even for one channel or platform, is daunting enough for many brands. Doing this at scale is another challenge entirely, though. Ultimately, you have to apply this to every channel, across the whole consumer cycle, to get the depth of insights that maximise performance in short time frames.

Pulling data from multiple channels gives you a deeper understanding of your customers, but also how these channels pull together throughout the consumer cycle.

The role of these channels and the effectiveness of different creatives drastically changes throughout this cycle, too. Peak performance at the top of the funnel is wildly different from the bottom of the funnel, and the “messy middle” in between is only growing.

Need help with applying data to creative?

Our Performance Creative team is ready to help you scale up your advertising strategy by applying data to every stage of creative production. Drop us your details, and one of our team will get back to you, or you can learn more about our Performance Creative services here.