Written by Roman Stanek |
As I wrote in my blog post for the GoodData Spectrum launch, we’re at the point when every business critical decision and every single step within the decision-making process can be made and supported by data. This reality expands the world of BI and analytics from a select few—like business analysts—to a far greater number of employees in a company, all of whom benefit from access to data to make critical business decisions. In a future where the use of analytics is truly pervasive, multi-tenant models will be necessary to make the insights personalized and contextualized.
I’m reminded of when Salesforce first edged out Siebel, thanks in large part to its multi-tenant architecture. Initially, Salesforce was derided for its multi-tenancy, with commenters claiming any mishap would lead to a “negative domino effect” among other users. These fears were, of course, unfounded, with people assuming only a few large companies would need CRM—which we know now is not the case. Essentially, people didn’t believe that Salesforce’s use of multi-tenant architecture and the low cost of management associated with it was necessary, because they didn’t believe that its use would be widespread enough to merit it. Salesforce proved the value of putting data in the average user’s hands, and it changed the game for SaaS as we know it.
Yet in the world of BI, people are still having a discussion about the merits of multi tenancy, because the perception is that BI is unique and somehow exclusive to C-suite employees or data scientists. Many on-prem vendors now offer support for cloud-based instances. And while spinning up an instance on AWS is definitely faster than procuring an on-prem infrastructure, the complexity of management, monitoring, securing, upgrades and performance tuning is exactly the same in the single-instance cloud as it is in on-prem solution.
Now imagine that you want to use BI for every single process, department, line of business, and customer across the entire spectrum of decision making. There’s much to consider. For BI to be business critical you need to assure uptime, performance, concurrency and many more enterprise capabilities. With native multi-tenant platforms, each user’s experience can also be managed (and I am using the word “manage” to describe everything from installation, upgrades all the way to security and GDPR compliance) efficiently from one central point—instead of managing each separate instance from a different console. Multi-tenancy is the only option scalable enough to extend the pervasive experience to thousands of users each with their own context and requirements.
For me, your view on the need for multi-tenancy comes down to where you see analytics going. Do you see analytics becoming more pervasive and democratized, used to improve business processes throughout the organization so companies can operate faster and more efficiently? If so, which I believe it will, then multi-tenancy is a must.
Written by Roman Stanek |