Determining Granular Creative Aspects of an Online Campaign

Many years ago I was invited to produce an article for the now defunct 'Digitrends' magazine and their online presence 'Digitrends.net'. I took a little bit of perusing (again in the wayback machine), but I was able to find it. I've replicated it here knowing that I won't have to search high and low for it in the future.

Replicated via reference to this listing at the wayback machine


Determining Granular Creative Aspects of an Online Campaign

by Stuart M. Davis
Director of the Marketing Optimization Group, Beyond Interactive
(posted 01/03/2001)


From Digitrends, November 2000

The digital advertising media buyer is sometimes considered a “jack of all trades,” the person who needs to know how all campaign variables come together to create an effective buy.

But most aren’t graphic designers with art school backgrounds, and the others may not be masters at data collection and analysis. So how does the media buyer determine how the granular creative aspects of a campaign correlate with site performance?

For example, how does a buyer know that blue banners get clicked more than red ones on a particular site? Or where creative that incorporates a specific call to action does perform best on a page? How do media buyers compile and use this knowledge to adjust campaigns in real time, anticipate future performance, and model new campaigns?

First, we need to define the role of the media buyer. A buyer should focus on buying, not design and analytics. Building a relationship with publishers and vendors is a job that requires full dedication. It’s also a job that’s essential to the creation of a successful media buy.

This is not to say that a buyer shouldn’t have a well thought-out set of data to back up his ventures. At Beyond Interactive, the marketing optimization group is the box of knowledge our media planners and buyers use to get the data they need about creative and site performance. The numerical experts in this department constantly analyze campaign data to uncover correlations and explore tactics that help campaigns to perform better. By pairing together strong communication and leading edge technology, the marketing optimization specialist can guide the media buyer down the right path with the right data to make better and more effective deals.

The Tweaking Never Stops

This share of knowledge definitely does not end with the media buyer. The digital industry allows advertisers to tweak campaign placements and creative in real time. So how does one utilize this opportunity to the fullest potential? Marketing optimization.

We’ve always had the ability to optimize creative placements. When a creative unit is under-performing, that creative is removed from rotation. The problem, of course, is to understand at what point to actually pull that creative. Have there been enough impressions to justify the performance? Have there been enough impressions with other creative to really compare them? What if the rest of the creative start to under-perform? Is there an aspect of the creative that could be changed to enhance performance? One part of the answer to all of these questions is to make sure that all decisions are made with statistical analysis.

Another part of the answer is to understand that optimization is not just about removing creative from rotation. The opportunity also exists to re-apply a creative to the placement. The technologies that exist in Beyond Interactive’s Lifeline allow the optimizer to know that a creative is under-performing at a statistically significant level. The technology also allows the optimizer to understand when a creative may no longer be statistically under-performing in reference to the rest of the creative on that placement.

Here’s an example. A piece of creative is cut from rotation because it’s performing significantly worse than other creative. Advertising waste for that placement is immediately cut down because we are no longer serving a poor piece of creative. Possibly due to burnout, the other creative begin to perform worse. Now it seems there’s no longer any statistical difference between the creative previously removed, “creative-a,” and the creative that are currently running. One cannot assume that “creative-a” would have performed in the same downward manner. Statistically, it seems that “creative-a” would perform in the same manner that it had when you removed it from rotation. So the only plan of action is to turn that creative back on. Or is it?

Any great creative designer wants to know that her designs are helping a campaign perform at its best potential. The designer may even be willing to help you make the campaign run better. However, a designer should never agree to a change unless there’s some historical reference that proves tweaking should occur. If only she knew what part of the designs to tweak. How does the creative designer know what actions to take in the middle of a campaign? Let us take a step back for a moment.

Tailor Accordingly

Everyone talks about the wealth of data that exists in digital marketing. Advertising success easily can be tracked with the proper technologies. The pieces of data that are often difficult to find are those that refer to the qualitative creative attributes that lie beyond simple traffic data. The Lifeline technology at Beyond Interactive injects quantitative data into these variables in order to draw conclusions and move campaign optimization a gigantic leap forward.

There has been a lot of emphasis on the tactical process of removing creative from and adding creative to the rotation for any particular placement. Unfortunately, this tactic turns an artistic execution into a mere piece of media. When you start to consider the variables within that piece of art, the creative designer can take those variable findings and execute adjustments to the creative currently running. The idea is to develop the most successful creative execution available.

Now we can return to our example. Amazingly, the marketing optimization group informs the creative department that statistics for the current campaign show that the site’s target audience, 21-year-old males, respond better to creative with green backgrounds and creative with photo-realistic images of bananas. None of the existing creative contains both of theses variables together. The variables are now pieced together and pushed out to the campaign instantly. So instead of placing a previously removed creative into the rotation, a new, statistically optimized piece of creative that has incorporated the best of the attributes available is set in motion.

Creative executions now have a chance to improve instead of being subjected to simple light-switch style optimization. Media placements get to benefit from a tailored piece of creative that can relate to that placement’s design and audience. If one site seems to be less effective than another, we’re no longer forced to assume that the problem resides with the site. It may be that creative running on that poorer performing site just don’t match the response styles of the site’s audience. Those that use the tactics and systems mentioned will have the opportunity to test multiple iterations of a creative concept to determine the true reasons for why a placement works or does not work.

With all of these tactics and analyses, the campaign comes full circle. The media buyer and creative designer are much more powerful. “Knowledge is power,” and the marketing optimization group makes sure the collected data turns into relevant knowledge and that the power is translated and handed to all those who can benefit from it. The media buyer now has the data and hence knowledge to make educated buys. The creative designer now knows what attributes translate into efficient responses.

There’s plenty of technology in the world today to help the advertiser find data points and organize data into comprehensive reports. Companies interested in using the best business practices must use all of the technologies at their grasp while confirming that they don’t rely entirely on that technology. There must be experts to use the technology and proper channels of communication to distribute that knowledge. At Beyond Interactive, the marketing optimization group uses proprietary Lifeline technologies to collect data and aid in the analysis of that data. The optimizer then takes that data and all of the relevant knowledge associated with that data and molds it into understandable communication tools for media buyers, creative designers, and other beneficiaries at Beyond.

Stuart M. Davis is director of the marketing optimization group at Beyond Interactive, online advertising consulting firm based in Ann Arbor, Michigan.