The acquisitions will provide Google and Salesforce customers more capabilities to analyze and create visualizations from data, helping to beef up analytic offerings to better compete with Amazon and Microsoft.
While these acquisitions serve cloud companies’ desire to provide more services and tools to build analytics, they miss a massive opportunity to change the game. The real challenge cloud platforms should be addressing is not just adding one more end-point analyst capability, but in connecting all of these solutions or applications to a trusted analytic management foundation to operationalize analytics in a highly consistent and efficient way.
Today, each time an analytic calculation is created for a visualization (typically within SQL or in the application itself) it becomes locked within the application, limiting transparency and governance while also adding maintenance complexity and risk. That might seem fine on the surface for an individual analyst, but it creates an enterprise-scale problem when other analyst or data science teams need to replicate this logic in the context of machine learning models or operational applications. Leading enterprises increasingly recognize that these analyst activities are not independent of each other but must be reconciled and consistent to promote trust and ensure auditability/repeatability.
The data science teams are often charged with taking a descriptive analytic (how many purchases were made in the XX category) and creating a predictive view (how likely is a customer to make a purchase in that XX category in the next 60 days). This leads to recoding and more data munging in their tools in an attempt to have consistent features or dependent variables matching the results of what the business sees.
If these analytics do not match, business risk is introduced. Companies see this all the time when their promising prototypes in the data science test environment don’t deliver the same results when deployed in production.
The market needs a solution to make analytics more consistent and reusable across individuals, teams and enterprises—a “place” where analytics are accessible by the business, business intelligence and data science teams. This requires a shift in mindset to manage not just the data but also the math from the very beginning—the first time a line of code is written.
Imagine a world where your analysts and data scientists had access to the answer or solution to every previously asked business question and where you had high trust in every analytic your organization produced. A world where you could spend more time innovating and less time re-creating, reconciling and worrying about analytic differences.
This is all possible with a trusted analytic layer, and it is going to be the next big investment area where companies and investors will realize exponential gains.