For Next-Level Campaigns, Leverage the Power of Prediction

AI can pick up where dynamic creative optimization leaves off

Connect, convince and convert: I believe these three actions comprise the holy trinity of an optimal consumer journey. And while plenty of offerings address the connecting and convincing stages of the consumer journey, few options exist for advertisers focused on conversion. Solutions include retargeting, SEM and dynamic creative optimization (DCO).

The latter has become a particularly hot commodity due to the rising demand for personalization coupled with advancements in machine learning. The surge in inter­­est around DCO comes from advertisers keen to adopt technology that allows them to personalize ads based on user context. Despite DCO’s benefits, challenges remain.

One such challenge comes in the form of pre-set rules. Pre-set triggers and decisioning enable DCO to validate existing assumptive approaches, but can be limited. For example, you may want to serve visitors to your website more ads when it’s raining. DCO stipulates that you set a particular message for those conditions and rely on technology to prescriptively serve that ad when rain occurs. If the ad performs better in those settings, then your assumption is correct. But if the results show otherwise, you will likely need to test a different hypothesis through another campaign.

Even with the infusion of AI into DCO, there can still be difficulties learning and self-adjusting over time as various creative elements are dynamically assembled. You’d need to evaluate a campaign’s performance following its flight and adjust the creative elements for the next round based on whether your assumptions proved to be true.

A third challenge stems from the technology’s contextual dependence. It relies on consumer context to operate, meaning that in the absence of such context, DCO may not be as effective. And as cookie-based targeting wanes, it might soon become even more difficult to rely on that all-important context.

Searching for solutions

Advertisers should be able to expect more from their digital campaigns beyond DCO’s limitations. Ideal solutions should harness the benefits of personalization without forcing a long lead time. Media buyers and brands should look for solutions that improve the effectiveness of their advertising campaigns in the following ways:

Anticipation: A solution that predicts the right arrangement of creative elements can drive outcomes by audience. Anticipation can help drive the highest engagement, conversion or both for a given audience, prior to your ad being served. While DCO requires user context, a solution with anticipation can better respond with the right message by deconstructing and reassembling creative assets on a highly granular level. This process allows for more creative flexibility and goes further than DCO does in helping you predict what messages to send to consumers.

Segmentation: By using a tool that automates the discovery and segmentation of audiences based on message resonance, you can uncover new audiences with greater speed and effectiveness than DCO tends to allow. A solution that incorporates in-market data processing, model creation and training can help identify hidden engagement patterns across audiences. You can create clusters of users who share similar characteristics and behaviors and optimize your creative approaches for each one.

Revelation: By analyzing each campaign from a creative and audience perspective you can extract rich insights around the highest and lowest performing experiences. This represents a unique opportunity to inform future strategy by measuring the impact of your campaigns among target audiences.

A problem fit for AI

Fortunately, smart loves problems (as we say at IBM), and in analyzing the inherent challenges associated with DCO, we saw an opportunity to leverage IBM Watson AI to empower advertisers to stop reacting and start predicting. This power of prediction is the bedrock of IBM Watson Advertising’s newest solution, IBM Advertising Accelerator with Watson.

Leveraging some of the latest advancements in Watson technology, Accelerator operates on a predictive model to help enable media buyers and brands to be more flexible and granular in their creative approach. It can also enable them to find the optimal combination of digital display elements to help drive the highest engagement and conversion for a given audience.  Accelerator also segments audiences and returns insights about what creative elements drove the best response for a given segment.

To drive installations for The Weather Channel’s Storm Radar app and show how the offering delivers on its promise, we tested the technology on ourselves. After only 23 days of using Accelerator, we saw a dramatic performance lift of 3X in installs.

Given the great success we saw in our own campaigns, we’re excited to introduce this offering and help enable more marketers to accelerate their advertising.


Dave Neway is the head of marketing at IBM Watson Advertising. He is responsible for ideating and executing the go-to-market strategy for all Watson Advertising offerings. Prior to Watson Advertising, Neway was a director of strategic solutions at Pandora, where he achieved double-digit revenue growth within the CPG and entertainment verticals.