A Primer on AI-Powered Creative Analysis
by Isabel Greenfield9 min read
- Creative analysis is the use of artificial intelligence to evaluate creative elements within ad creative to determine their influence on viewer action.
- By analyzing large amounts of video creative and performance data, the relationship between different creative elements and specific performance outcomes can be identified.
- Creative analysis can examine elements such as the number of people on screen, tone of voice, theme, and more.
- With the use of creative analysis, advertisers can streamline the production process of their video ads and maximize the return on investment. Creative analysis can also help brands make their foray into other platforms like Instagram and TikTok.
- The world of data collection is rapidly changing as people become more aware of how their personal information is being used and stored. As large companies like Apple and Google make changes to their policies, advertisers are finding new ways to reach their target audience and maximize their ad budget on channels like social media.
What is Creative Analysis?
Creative analysis is the use of artificial intelligence to analyze video assets and determine what specific creative elements (such as a voice over, number of actors, theme, etc.) are most effective at delivering specific outcomes. This can be conducted on any ad channel that has a video ad unit—from social media, to Connected TV.
To see an example of the insights creative analysis can provide, see our What Makes the Perfect CTV Ad? report.
Creative analysis is a type of computer vision, a form of A.I. that teaches computers to “see” differences in visuals and analyze the effect of these differences. Advertisers are beginning to use this technology to analyze their video assets and make strategic creative decisions based on the findings. And this can’t come at a more opportune time.
Historically, audience targeting has been advertisers’ focus, with ad spend being used to test who is being reached rather than what creative they are being reached with. However, as data privacy has come under consideration on mobile and social media over the past few years, marketers will have less and less audience data to use for their targeting efforts.
With this lever becoming less important, the creative takes centerstage. Creative analysis uses A.I. to go beyond the standard A/B test, giving marketers valuable information about what is really moving the needle for viewers. It can hone in on subtle aspects of video creative, from tone of voice to number of people on screen, to determine what viewers respond to while helping inform new creative decisions for advertisers.
How it Works
- At the core of creative analysis is computer vision, the technology that allows a computer to teach itself to “see” the difference in images.
- Computer vision uses an algorithm vs. a manual, human upload.
- By taking in a large amount of images or audio tied to the same subject, a computer can start to spot the differences between them.
AI in Advertising:
- The more data the computer is able to process, the more accurately it is able to understand the information. This is why it is beneficial to use the technology at scale.
- When it comes to advertising, the more video assets that the AI-driven creative analysis processes, the better.
- This provides more opportunities to analyze the varying elements of the videos more clearly. The more data it has, the more accurate (read: actionable) it becomes.
Creative Analysis and Measurement:
- Pairing performance data to each video asset analyzed is the key to connecting which creative element excels at delivering a specific outcome.
- Equipped with this data, creative analysis technology can draw a conclusion.
- For example, it can recognize that videos utilizing a voiceover have a better chance to drive more sales than those that do not.
What Exactly is Computer Vision?
Computer Vision is a type of artificial intelligence that trains computers to collect data based on visual inputs, in a sense allowing a machine to “see.” It does this by feeding the computer large amounts of data, allowing it to begin to tell the differences between various images. Rather than a human feeding this information to the computer, it uses an algorithm to continue to learn. This is the type of technology that allows self-driving cars to “see” the road while they are driving.
So why are we talking about this here? Creative analysis is another type of computer vision that allows a computer to see the difference between ads and analyze the impact those minor changes have on performance.
What Can Creative Analysis Tell Advertisers About Their Ads?
Creative analysis goes beyond an A/B test that only looks at one variable such as call to action or background color. As we mentioned above, computer vision can take in vast amounts of information and see slight differences, which comes in handy when getting granular about what is actually driving performance within an ad. Examples of elements that can be evaluated include:
- How many people are on screen at once
- How many of the actors have speaking roles.
- The tone or prevalent emotion used in the ad.
- The amount of time a product is shown on screen.
Creative analysis can look at all of these and determine what is most important for brands, so they can make educated creative decisions and ensure their ads work effectively to generate the outcomes they’re after.
What are the Benefits of Using Creative Analysis?
The biggest benefit of creative analysis is pretty clear—a greater return on investment for advertisers’ video assets. With insights into the specifics of the creative, advertisers will know what is driving performance and can make sure their creative budget is used effectively.
With creative analysis advertisers can make forays into other video platforms, such as Instagram or TikTok, using creative they already have. They can then learn exactly how the creative needs to be tweaked to be successful on that specific platform. Overall, advertisers’ video production process will be streamlined because they’ll know what changes they need to make, rather than making an educated guess or making multiple creatives to A/B test one variable.
The Shifting (Data) Landscape
User privacy has been a hot topic as of late and one with a big impact for advertisers. The collection and use of personal data are being reexamined, creating shifts in what data is being collected and what is available to advertisers for targeting. Apple and Google are making big policy adjustments to how they collect data, bringing this long-pending shift into reality.
Apple: Apple’s two recent software updates have radically affected audience targeting for advertisers. Those iPhone pop-ups are more than just an annoying reminder to update the software, they are actually radically changing the landscape of the advertising industry.
- iOS 14.5: App Tracking Transparency (ATT). With ATT, apps are now required to get a user’s permission before collecting data that will be shared with other companies for the purposes of tracking usage of apps and websites. This is a big shift from previous iterations where a user would have to opt out of sharing this information manually. When a user opts not to share their data across apps, they are limiting the amount of behavioral data that is collected and advertisers no longer can track their subsequent actions, making it harder to gather conversion information. This means behavioral targeting data (ex. “Pop culture fan” or “In the market for a car”) is harder to collect.
- iOS 15: Intelligent Tracking Prevention, Mail Privacy Protection and App Privacy Report.
- Intelligent Tracking Prevention allows users to block their IP addresses from trackers on the Safari app.
- Mail Privacy Protection allows them to block their IP address in the Mail app, limiting email senders from tracking engagement metrics.
- They also introduced a new App Privacy Report so users can see exactly how their information is being used if they do choose to share their data.
Google: Apple isn’t the only one making significant changes. Google is working to phase out the use of third-party cookies across desktop and mobile devices. A third-party cookie tracks a user across multiple websites, collecting data that establishes a user persona, which is used in many of the third-party targeting audiences that advertisers have come to depend on. Eliminating this use of third-party cookies will contribute to the lack of third-party data that Apple’s operating system updates have already started to limit.
What Does This Mean for Social Advertisers?
The shifting of data collection and privacy rules means that social advertisers are not going to have access to the same amount of third-party data as they have in the past. Over the previous years, audience targeting has been one of the biggest levers for advertisers to use to maximize their ad budget on display and social media. The creative they were using was often a distant second when it came to optimizing their ad strategy. With the diminishing amount of targeting information, advertisers will now have to shift their focus to make sure that their social creative is doing the heavy lifting, leaning on contextual targeting (more on that in a minute) to make sure their ad is reaching their correct audience and that it’s driving conversions when it is seen.
As the amount of third-party data decreases due to changing privacy rules, the demographic and behavioral targeting that is recently popular on social will become less effective. Instead, advertisers can turn to contextual targeting, meaning reaching audiences around content that aligns with what is being advertised. For example, a CPG company may choose to run their ads within cooking websites as people who seek out this content probably enjoy cooking and will likely be grocery shoppers. In the past, they may have used third-party data to target “grocery shoppers” which may have been gathered by location data, membership rewards, or app usage. Creative analysis will make sure that when social advertisers do reach their target audience in contextually relevant content, it will be as effective as possible.
What the Industry is Saying
“Marketers are looking to predict a hit and avoid a disaster before they make a significant investment.”Elizabeth Paul, Chief Strategy Officer at The Martin Agency
“Once largely considered an afterthought, the ads themselves are becoming as much a priority as the placement of them for some online advertisers — many of whom are testing new ways to work creative assets harder… ‘No matter how good you get at optimizing your media performance, if your creative is not fit for purpose then it’s not going to work as well — we reached that point of diminishing returns.”Ander Lopez Ochoa, Head of Digital, media and e-commerce marketing at Johnson & Johnson
“About 82% of survey respondents shared that they have shifted campaign creatives based on what creative ads their competitors have leveraged in the market, creating an inefficiency in terms of how this data on competitors is getting to CMOs.”BrandTotal CMO Study
“Recent data from McKinsey revealed that artificial intelligence (AI) has the potential to create $1.4 to $2.6 trillion of value in marketing and sales across the world’s businesses. What’s more, according to a survey of 341 marketers at for-profit U.S. companies by The CMO Survey, the top uses of AI in marketing include content personalization, predictive analytics for customer insights and targeting decisions.”Forbes
“Context, in terms of a viewer being in a frame of mind that is receptive to the message in question, is a regular topic in advertising. Its importance was further reinforced in a recent Integral Ad Science study showing that context can increase the memorability of an ad by up to 40%.”AdExchanger
Recommended Reading on This Subject
- Three Ways Computer Vision is Transforming Marketing
- 8 Marketing Trends You Need to Know in 2022
- Quickframe Creative Analytics White Paper
- WTF is Creative Analytics
Creative analysis can reveal which elements in an ad can help prompt specific outcomes for advertisers. This data can help inform future creative approaches, where advertisers include specific elements in their ads to ensure their goals are met. And as data privacy restrictions continue to take hold of popular video ad channels like social media, this technology can help advertisers develop intelligent creative to ensure they still achieve the impact they are after.
Subscribe to the MNTN Research Weekly.
Sign up to receive a weekly feed of curated research, sent straight to your inbox.