tableau_or_einstein
Dave Dixon

Dave Dixon

Tableau or Einstein: The Case for Both

It has been almost 8 months since Salesforce announced its plans to acquire Tableau, and 3 months since the acquisition cleared European anti-trust reviews allowing the companies to officially merge. While Salesforce is still working through what the long-term product roadmap will look like, one thing is already clear: In the near-term, both Tableau and Einstein Analytics will continue to be supported and developed.

Given the similarities and overlap in the product functionality, a question I often hear is, “Which tool should we use?” The answer depends on the specific use cases involved, but I believe that for many organizations the right answer is both.

As noted in my blog post written shortly after the acquisition was announced, while there is significant overlap in Einstein Analytics and Tableau, there are gaps in both offerings that are filled by the other. A few important distinctions between the tools include:

  • Tableau leads in “on-premise” analytics, commonly used for back-office and mid-office focused solutions.
  • Tableau allows for easy connectivity to common on-premise data sources without having to build complex integrations.
  • Tableau leads in real-time analytics, with the ability to directly query real-time transactional datastores.
  • Tableau lacks true AI/ML capabilities, like the kind available in Einstein Analytics Plus through its Discovery functionality.
  • Einstein Analytics was developed to work seamlessly with the Salesforce platform. Tableau can be integrated as well, but was not specifically developed to exist within the Salesforce user experience, and requires some customization to do so.
  • Tableau lacks the Salesforce action framework found in Einstein Analytics.

So, given the differences, which one should you choose? I think it’s important to approach this question within the framework of achieving an Intelligent Experience.

One of the cornerstones of the Intelligent Experience is surfacing insights where they can have the most business impact. This means providing insights within the flow of business processes where users are working and making decisions. This often means providing analytics and predictions within systems where users are already working. Depending on the users involved and the focus of the analytics solution, this could mean surfacing insights in different platforms for different users.

For instance, sales reps and sales managers typically rely heavily on Salesforce for opportunity and pipeline management, and use Salesforce’s mobile application for accessing their CRM “on the go.” Given Einstein Analytics’ ability to seamlessly embed in both web and mobile Salesforce interfaces, and the ease with which insights can be made ‘actionable’ via the Action Framework, it makes sense to surface opportunity, pipeline, and forecasting insights in Einstein. However, back-office users such as supply chain or logistics managers may not use Salesforce. Additionally, these users may need real-time updates on stock and inventory levels – use cases for which Tableau is a better choice. So which analytics platform should we choose? The reality is that diverse use cases mean that different users are best served with different analytics tools. In this case, we’d use Einstein for the sales team, and Tableau for the supply chain and logistics groups.

One objection to this approach is the risk of ending up with “different versions of the truth,” especially when subject areas overlap between use cases. However, this issue can be addressed with sound data architecture practices. Having a well-curated, central repository for master data elements and key shared subject areas can help ensure insights reflect a “single version of the truth” regardless of the analytics tool used to surface them. For instance, in the example previously given, both back-office users, as well as the sales team, may have need of product hierarchy from a “product master.” Storing a single product master in an MDM or EDW and surfacing it in both Einstein and Tableau ensures a consistent view of the product hierarchy. By moving data curation and business logic to a shared data layer, we can provide the ability to use whatever analytics solution works best for the user groups involved.

Another objection to using multiple analytics solutions has been the effort involved in maintaining and administering multiple analytics environments. With cloud-based analytics solutions such as Einstein Analytics and Tableau Online, much less setup and administrative burden are required, alleviating this concern.

Given the decreased burden in supporting cloud analytics solutions and the clear benefit of surfacing insights in the best platform for the use case, multi-product analytics solutions are often the right choice for an enterprise. If Tableau and Einstein are both pulling from the same shared data repository, then there is much less risk of having “different answers in different systems.” By tackling the data consistency concerns with solid data architecture, we can focus our attention on providing users with the best experience possible, using the tool most tailored for that experience. For analytics solution success, the user experience is just as important as the intelligence delivered. This concept is critical for ensuring user adoption and ultimately, ensuring solution effectiveness. For these reasons, many enterprises will find that using Tableau and Einstein Analytics together, supported by a well-designed data architecture, will provide better outcomes than leveraging a single tool alone.

Related Content

Share this post

LinkedIn
Twitter
Facebook