IBM Watson Advertising: IBM Advertising Accelerator with Watson


Campaign Summary

IBM Watson Advertising used its AI technology to promote installations of IBM's Storm Radar app.


Objective and Context:

In January 2020, IBM Watson Advertising wondered what kind of assistance AI could lend to the creative process. Its answer was IBM Advertising Accelerator with Watson ("Accelerator" for short). Accelerator rapidly and continuously learns which creative elements will resonate with each audience based on not only how consumers react but also on a multitude of other key signals such as DMA, device type, and time of day. It then delivers a unique ad unit loaded with the most inviting creative to the right audience. Moreover, it was designed to get better and better every time, just like the human brain.

IBM Watson Advertising tested IBM Advertising Accelerator with Watson on its owned and operated Storm Radar App to increase app installations and prove that this new offering could deliver on its promise.

Target Audience:

IBM Watson Advertising targeted current The Weather Channel app users on iPhone and Android.

Using IBM Advertising Accelerator with Watson enables the automatic optimization of creative and audiences during campaign flights through AI and machine learning. Advanced Watson AI rapidly learns and then predicts the optimal mix of creative elements to help drive engagement. Accelerator uses predefined creative assets that can be deconstructed and re-assembled on a highly granular level, providing advertisers with exponentially more flexibility in their creative approach.

The creative generated by Accelerator is personalized. Each time an ad is served, it is a unique creative to each performance group and designed to maximize resonance. Using Watson AutoAI, Accelerator helps automate the discovery and segmentation of audiences based on message resonance. As the first IBM offering to put Watson AutoAI to work, Accelerator enables faster in-market data processing, model-creation, and training — all of which help to identify hidden engagement patterns across audiences and to cluster similar users together.

Creative Strategy:

IBM Watson Advertising used Accelerator to predict the right combination of assets for each unique consumer. Accelerator identified hidden engagement patterns across audiences and continuously analyzed performance throughout the campaign so creative elements could be adjusted on the fly, driving operational efficiencies.


Overall Campaign Execution:

IBM Watson Advertising tested Accelerator on its owned and operated Storm Radar App. Using three different creative approaches and a control ad, Watson AI automatically optimized creative separately for iOS and Android, created user clusters with similar characteristics (i.e., created audiences), and optimized toward the top-performing creative for each audience.

Mobile Execution:

This Storm Radar app campaign was entirely mobile focused. The IBM Advertising Accelerator product is focused on optimizing creative elements for performance and efficiency on action-driven KPIs, such as those used to guide marketers' work in the mobile and app space.

Results (including context, evaluation, and market impact)

Utilizing three different creative approaches and a control ad, IBM Advertising Accelerator with Watson increased Storm Radar app installaations by over 3 times over a span of just 23 days.

Moreover, in the six months that Accelerator has been available, clients using it have seen at least a 25 percent lift in performance over the course of their campaigns.

Categories: | Industries: | Objectives: Innovation, Machine Learning and AI | Awards: NA Finalist, NA Silver Winner