Adobe Analytics: A Sneak Peek Inside the Labs
At Adobe, we believe that AI and machine learning technologies are most valuable when it can amplify human intelligence. As the modern-day business environment becomes increasingly challenging, this will become a key element for brands to build competitive advantage. AI will ultimately give workers the tools to automate cumbersome processes, uncover hidden insights within their data, and tackle complex questions with much more precision.
In Adobe Analytics, we have secretly been working on new technologies, powered by AI and machine learning, to help users gain a deeper understanding of what is happening within their businesses. These forward-looking innovations will help us continue moving Adobe Analytics clients beyond simple vanity metrics, tapping deep data insights to power every part of the experience they deliver to customers. We are pulling back the curtains today, giving a sneak peek into some of the projects the team has been working on.
When forecasting future orders and sales, the real value comes from the ability to address anticipated shortfalls. This new Adobe Analytics project delivers on this, giving users the ability to forecast against business targets, while leveraging AI and machine learning to recommend specific actions that can help close the gap between targets and projections. Billions of historical and real-time data points are analyzed in this case, pinpointing a set of tactics that will help optimize the business. Take an online retailer for instance, who is forecasting a shortfall in the Summer months. The system could recommend driving 5% more conversion amongst the top 3% of spenders, an event that would close the gap faster than targeting the full customer base. In a different example, an online travel company forecasting a slower Fall season could see that optimizing the mobile experience and driving 10% more traffic would help the business meet their targets.
Data can either help brands enhance the customer experience, or fix broken ones. When it comes to the latter, the urgency is greater but the path forward is not always clear. This project in Adobe Analytics helps address this, allowing users to take an undesired event and hone in on the primary contributing factor. Take a media company for instance, that is seeing irregular spikes in app uninstalls on smartphones. The system will begin by analyzing all the customer “paths” that are relevant for the brand, looking at how users navigate across different properties and devices. It will begin to narrow this down, identifying the top contributing paths that lead to app uninstalls, and allowing the system to make a recommendation on the top contributing factor (or factors). The brand could see for instance, that a key feature is broken and causing user frustration and fallout. Or, there may be a disconnect between the app and desktop Web experience that is pushing users to pick one over the other.
No matter what the industry, all brands struggle to identify their most valuable customer segments. There is a dizzying array of factors to juggle, from different age groups and demographics, to various purchase patterns and preferences. Unfortunately, much of the work around building customer segments is left to guesswork, creating challenges when it comes to properly engaging new customers and nurturing existing ones. This Adobe Analytics project takes guessing out of the equation, tapping AI to deliver much greater precision. The system will go through and crunch billions of data points that are relevant to a brand, looking at user profiles as well as known preferences and behaviors. The most valuable groups of customers will be automatically clustered and presented. An online retailer for instance, instantly has access to all the most relevant customers to engage for trying a new feature in their app, or for an upcoming promotional campaign. They may also find unexpected surprises, such as a crop of users in a particular region that converts at higher rates.
Analysis workspace assistant
Within organizations that number in the thousands or tens-of-thousands, analytics users often make very similar queries. One business team might feel they have a unique and complex question to answer, while a different team had already cracked the code. A lot of time is wasted in enterprises through replicated efforts. This project in Adobe Analytics aims to address that with an AI assistant. Take a team that is trying to better understand recent fluctuations in revenue. For questions such as this, there are a lot of different areas of the business to analyze, and a great time of time is spent figuring out the starting points. The assistant however, will take this query and begin by assessing all past queries that have been made across the organization—looking at similar ones to suggest initial areas for the user to examine. One suggestion could be to break down revenue by browser type, leading the team to a conclusion that a subpar experience on Internet Explorer is causing substantial user fallout and revenue loss. As users make queries over time, the assistant will learn which recommendations provided the most value and further refine future suggestions each time.
Learn how Adobe Analytics can help you understand how customers navigate through their entire journey.