Clouds Everywhere - Data, Applications, Analytics, Services

 As the New Year descends upon us, I’m excited to see that the promise of cloud technology has come to fruition.  What am I basing that on? Well, recently, a doctor friend of mine commented about how she could easily solve her problem by storing data in the cloud..  Don’t you just love it when a complex technology environment can be summarized in a single, eight word comment?  For me, it means that the use cases and the technology have evolved to a point where mass adoption is occurring.  Technology vendors call that safe opportunity.  As clouds become ever more pervasive, what does that mean for the analytics that are driven from these “hosted environments?”  I see several trends emerging with the maturation of the cloud platforms and analytics.

Subscription analytics

For example, SalesForce.Com (SFDC) is only too happy to help with this new business opportunity of delivering analytics.  The idea here is that SFDC creates your analytics for you, and you simply consume the analytics in the cloud.  The benefits here are obvious – business users have a consistent view of the analytics, the analytics are readily available and stay up to date,, and the cost could be very reasonable.  As always though, with all good things there are limitations; in this case, SFDC creates the analytics, meaning the business is restricted to and dependent on the algorithms and analysis that SFDC found relevant and interesting. A second shortcoming to this model is that with greater adoption, there may be more interest, and as with anything else, more volume means more cost incurred by the customer. Finally, you don’t get to interact with the analysis as your own – you are only borrowing the analytics, similar to the way you’d borrow a book from the library.  Lastly, all of SFDC clients have access to the same analytics you are being served, making differentiation harder to come by. Many companies might prefer analytics delivered that are unique to their organization.

Hosted Analytics

The next option is to have people host your analytics for you.  This is subtly different from the subscription model here you rent space, the data, and the application to create these more customized analytics. Hosted analytics are very popular for many of the same reason as subscription analytics.  It has the added benefit though of enabling you to massage or mash the data in ways that are unique to your business. In addition, your own business analyst has access to this dedicated environment.  The downside, of course, is that the analytics live outside your firewall. In addition, the cost can be large if you have many users, and providers often don’t allow you to take the data inside your own firewall so if you want to combine it with your own corporate data. Instead, your corporate data has to be up loaded to the hosted site, opening up a sea of potential of governance issues.

On-premise analytics

The traditional path to analytics is one where companies create their own analytic warehouse with an analytic server and tools to support this environment.  This approach has many advantages, especially since the tools that are available on premise are powerful and inventive.  The downside tends to be cost (both in capital and maintenance) and agility.  Traditional on-premise analytic environments require hardware, software, and services.  The other downside revolves around supporting these systems – resources are required, support must be given in a timely fashion, and all this work falls on corporate IT, which is often times already overburdened with requests.  Business analysts have many requests to access, change, or deliver on data, reporting, or analytic requirements.  These tend to queue up and create latency and delivery problems in IT.  Last but not least, the data environment that is required to support complete and meaningful data is often housed in a myriad of places – i.e., many different data types and platforms.  Data that is found in a hosted application often cannot be brought down from the vendor to the client, and even if the data can be moved, synchronization and data lineage become big issues that drive the inaccuracy often found in analytics.

Hybrid Analytics –Future Vision

Some in the cloud, some in premise.  This seems to be the solution for the future.  As customers uncover and create these hybrid solutions, we will see how well the existing tools adapt to the changes.  I am excited to see what our customers will be doing with the combination of hosted and on-premise analytics, as it seems like a promising path to achieving operational excellence in the near future.


Next Steps

Read the new research report, “Analytics in the Cloud: A study conducted by Enterprise Management Associates” for more details on how your peers are using cloud-based analytics.