In our review of last week’s MarTech news, we highlight the beta launch of Google Signals empowering marketers with a cross-device view of the customer, Facebook’s engagement-boosting tool to make still image ads interactive, and Oracle’s new open-source framework for deploying AI models.
Google beta launches Google Signals to help create a single view of the customer across devices.
What it does
This new feature, still in beta testing, will help track and market to customers across devices.
Why it matters for brands
Customers today interact with brands across dozens of platforms and devices. Unifying those interactions in what we call omnichannel is the next stage of marketing and customer experience. This means more personalized ads and more increased conversations. Targeting ads across devices has been shown to increase conversions by around 40% more than just single-device marketing.
Why it matters for users
In a world of data breaches and privacy scares, the idea of being tracked across devices and browsers can be quite unnerving. Only users that have turned on Ads Personalization will be tracked. As always, be educated on what permissions you give. Brands, be transparent with your users.
Facebook boosts engagement with a new way to add lightweight motion to still image ads.
What it is
Facebook offers a new production framework called Create to Convert, which enables marketers and creatives to add basic motion such as brand elements or calls to action to still images. Brands using the tool have seen up to 5.5 times better conversation rates and 5.7 times lower cost per registration from video ads versus still ads.
Why it matters
Interactive content engages more than still content. However, video can be laborious and expensive. Simply adding interactive elements to still images is a cost-effective way for brands to boost engagement—driving up to 47% more time spent with the message and a 9x higher purchase intent.
Oracle releases GraphPipe, an open-source tool for deploying AI models.
What it is
Built by Oracle architects (one of whom developed at NASA), GraphPipe is a new framework that intends to fill a market gap between artificial intelligence and the applications that use it. It’s essentially a “standard, high-performance protocol” for transmitting data.
Why it matters
Especially for large organizations trying to implement AI into various capacities and departments, frameworks that are generic (from a provider) are often slow and inefficient, whereas custom frameworks (from the internal team) can be too weak for a growing bandwidth. GraphPipe supports five main already existing AI frameworks: TensorFlow, Microsoft Corp.’s CNTK, PyTorch, mxnet and caffe2. The code and developer resources are freely available on GitHub.
Before you go