How is Google building Cloud business?

Sundar Pichai Google
Meaghan McGrath, analyst at TBR, says Google is building Cloud business on corporate innovation to accelerate adoption.

While advertising revenues consistently dominate Alphabet’s revenues, Google’s cloud business continues to grow steadily. With global infrastructure expansion efforts under way and cloud portfolio feature expansion, particularly in the realm of machine learning, a continual focus of growing engineering teams, TBR estimates that Google’s cloud revenue (comprised of Google Apps for Work, for Education and for Government, as well as Google Cloud Platform revenue) grew almost 62 percent to $989 million in Q2 2016, and will continue to expand over the $1 billion mark in 3Q16.

Machine learning is core to product evolution and differentiation across Google

Sundar Pichai, CEO of Google, noted on the company’s earnings call that more than 100 teams at Google are using machine learning, and that machine learning technology will drive the company’s future. In addition to energy efficiency improvements realized from applying DeepMind algorithms to Google data center operations, Google Cloud Platform (GCP) also has grown its machine learning capabilities for users’ development substantially across 1H16.

At the forefront of theses innovation are Google’s multilingual Speech API and Cloud Natural Language API, which, in addition to its Vision API and Translate API, empower developers with predefined models that parse voice, text, and image inputs to extract deeper information about the subject. However, while powerful in their preconfiguration and expansiveness of supporting data, these models cannot be based on data specific to enterprise applications. To address this gap, Google also unveiled Cloud Machine Learning (Cloud ML), which allows users to build intelligent applications and predictive models using their own training data and thus return tailored, enterprise-specific results. Cloud ML integrates with other GCP products, can be supplemented with pretrained models and other data from the TensorFlow machine learning library, and automatically ingests data and continues learning from new inputs.

Google also introduced TensorFlow Processing Units (TPUs) in May, which are proprietary chip built specifically for accelerated machine learning processing and tailored for TensorFlow. Though just unveiled to users, these processing unit have been in use at Google for more than a year, and will continue to increase the performance of GCP machine learning services as they gain greater traction in the market.

Building functionality to promote Google Apps for Work usage in enterprise settings is an ongoing endeavor

In competition with other business productivity suites, Microsoft Office 365 in particular, Google continues to build out functionality that will increase the competitiveness and differentiation of Google Apps for Work as it works to further penetrate the large enterprise market. Recently launched features that add these capabilities include both Google Springboard and a new User Hub.

Google Springboard enables users to search through all applications simultaneously and also recommends relevant tasks and business information to users throughout their workday. The artificial intelligence-backed content recommendation and search capabilities will speed user navigation of files and add value as a type of workday assistant, catering to businesses’ needs. Once past the Early Adopter Program and made widely available, the instant accessibility of vital and current information will drive adoption of the Apps for Work portfolio by enterprises looking to improve efficiency and productivity, while having an easy-to-monitor daily guide of current activities.

Microsoft, however, offers Office Delve, a feature similar to Google Springboard that provides a dashboard of current projects and documents that are being produced by connected users and enables users to search all of their files. The similarities between Delve and Springboard mean that Google’s addition of this feature only supports competitiveness with Office, rather than elevating and differentiating Google’s productivity suite.

Conversely, the new User Hub, which was added in March, provides a landing page for users that shows all of the Google Apps that are turned-on for that user, in addition to any applications that have already been approved by the user’s admin, or that may be of interest and are available for download. While this feature does not necessarily improve a user’s daily efficiency, it provides user self-service and relieves some IT communication stress. Seemingly small feature additions like this can still elevate Google Apps for Work’s enterprise appeal.