Dynamic Creative Optimization

Dynamic creative optimization (DCO) has a rich history in ad tech. The basic principle behind DCO is that, instead of using a static ad unit, it should be customized to the user, site, geo, etc based on a robust set of rules and optimizations that will increase the overall engagement with the ad. This concept is largely agnostic to the medium, meaning DCO as a discipline applies to banner, native, video, audio, out-of-home, or anywhere that digital advertising can go. Some of the primary data points that DCO leverages include time, device, weather, context, location, behavioral information, demographic information, and retargeting. 


An important component of DCO is multivariate testing. This may include the data points listed above, or it may simply be optimizing a creative over time. For example, DCO may optimize copy, font size and color, call-to-action characteristics, and aspects of the creative itself, to simply find the most effective creative given various parameters (e.g. for a certain behavioral segment or within a specific contextual vertical). A very simple example might be an auto company - one that shows a creative in a rugged outdoors setting and another in a metropolitan context. DCO can ensure that the right creative is shown as often as possible to the right target audience.  

Some DCO providers allow a complex rule structure to determine what creatives elements to show based on what factors. Another strategy is machine learning to optimize on the full set of parameters at once. Both generally attempt to show the best creative to the right audience, then expand the audience that the creative works well for, for as broad a composite as it continues to work well. There is often a complex process of setting up the rules, business objectives, and measurement capabilities with the DCO vendors. As such, there are long sales cycles and the relationships tend to resemble SAAS vendors as opposed to media companies (though not always). 

Video is an emerging growth area for DCO. Companies like Sunday Sky create video ads that assemble a single video ad from a number of components based on a rule set that individualizes the overall video. The complexity is increased in video where there are challenges like transcoding a dynamic video to the right level of compression in real-time for the device. 

Have an independent DCO vendor creates an interesting dynamic for a programmatic campaign. This means that the DSP that powers that algorithmic buying has factors beyond its knowledge or control that effect the quality of the optimization. The DCO vendor may optimize in various contexts which may change the performance characteristics for inventory independent of the DSP's optimizations. There is generally no feedback mechanism between the DCO and the DSP for optimization. Because of this, media buyers thus may not be as aggressive in leveraging DSP optimizations when using DCO creative. Instead, the algorithm may be focused on buying the specific audiences or contexts for which the DCO may then serve the right personalized / optimized message.

Google is among the largest DCO vendors. The company purchased a leading DCO vendor, Teracent, in 2009, and has been actively developing the technology. This past week, Spongecell and Flashtalking - two relatively large DCO platforms announced their merger. In their press release, they noted "[i]ncreasingly [brands are] recognizing data driven creative is essential to connect with today’s consumer... Together we now deliver the market-leading DCO solution within the context of a complete buy-side ad serving solution: a truly independent technology and analytics partner without media buying conflict or bias, representing a clear and viable alternative to DoubleClick."  This makes it clear that a merger enables viable competition with Google's scale and product offering, as well as furthering their technology stack around data-driven creative. Other DCO players include Jivox, Sizmek, Criteo (via its retargeting platform), Adacado, Celtra, Adobe, and Rich Relevance. Some DSPs, like Rocketfuel (now Sizmek, but separate from Sizmek's DCO offering) have integrated DCO capabilities. At TripleLift, we partner with many of the largest DCO vendors to enable personalized and optimized native creative.

It is very likely that GDPR will have an adverse impact on DCO. DCO vendors are not in a position to ask for user consent, thus location - and information derived from location like weather, behavioral information, demographic information and retargeting will have limited use in Europe once GDPR becomes effective. That said, there remain several attributes of DCO that will continue to be effective, especially around generally optimizing the creative.