Back in 1999 (when the moon did not have a moonbase Alpha, nor did an explosion send the moon rocketing across the cosmos -- a reference for old-timers like me) while at my last startup, Bluestreak, we started experimenting with dynamic creative.
The idea was that there were e-commerce companies with thousands of products available online, and that based on location we should be able to test and optimize which products led to the most clicks and purchases. Over the next few years, we worked with several customers to experiment with this. We ultimately ran ads with several publishers that would rotate through a list of products, and we used our creative optimization technology to determine which combination of offers was getting the best results (based on clicks, interactions, or conversions).
It turned out that there were various combinations of location (publisher) and product that worked much better than others, and the tests were successful. But the question was really about matters of degrees. We saw significant improvements in results, and we developed great technology that supported all this. But after the bottom dropped out of the market in 2000 and 2001 and the price of inventory dropped significantly, the improvements in performance stopped mattering as much.
Essentially, the price of inventory was so low that it was cheaper to just run much higher volumes of unoptimized ads than to pay for optimization service.
But I knew that creative optimization and dynamic creative would have its time and place. Either the impact of the creative optimization would drive significantly better results, the price of inventory would come back up, or we'd be able to optimize the offers based on user targeting rather than just by publisher.
Creative optimization and dynamic creative dropped out of the industry for eight to 10 years, but it came screaming back. As I guessed, the major driver was targeting based on user data. And over the past few years, the growth of real-time bidding and audience targeting has led to significant improvements in dynamic creative and optimization.
There are now several significant companies that have built their business around the idea of optimizing the offer shown to users based on their profiles, including a lot of retargeting. They build advertising campaigns that are driven by databases -- ones that pull together the creative in ways that include hundreds or thousands or even millions of possible combinations. The best offer is selected based on a variety of criteria, including audience targeting attributes such demographics, behavioral data, and retargeting data. This information is extensively available and can be used to drive significantly better optimization than just location.
We all know that with real-time bidding and ad exchanges, ads can be targeted based on this kind of data. And we all know that with basic tracking of impressions, clicks, and conversions, bid prices in ad exchanges can be adjusted to optimize results based on the number of clicks or conversions. But dynamic creative optimization can take things to the next level. Using all of these technologies and techniques in combination can significantly drive up ROI. The only question is how many different products, offers, or options are available for optimization purposes.
The more opportunities to adjust the creative -- especially if those products or offers can be somehow predicted to match against different audiences' preferences or interests -- the more likely the user is to act.