BI In The Cloud? Yes, And On The Ground, Too

Slowly but surely, with lots of criticism and skepticism, the business intelligence (BI) software-as-a-service (SaaS) market is gaining ground. It's a road full of peril — at least two BI SaaS startups have failed this year — but what software market segment has not seen its share of failures? Although I do not see a stampede to replace traditional BI applications with SaaS alternatives in the near future, BI SaaS does have a few legitimate use cases even today, such as complementary BI, in coexistence with traditional BI, BI workspaces, and BI for small and some midsize businesses. 

In our latest BI SaaS research report we recommend the following structured approach to see if BI SaaS is right for you and if you are ready for BI SaaS:

  1. Map your BI requirements and IT culture to one of five BI SaaS use cases
  2. Evaluate and consider scenarios where BI SaaS may be a right or wrong fit for you
  3. Select the BI SaaS vendor that fits your business, technical, and operational requirements, including your tolerance for risk

First we identified 5 following BI SaaS use cases.

  1. Coexistence case: on-premises BI complemented with SaaS BI in enterprises
  2. SaaS-centric case in enterprises: main BI application in enterprises committed to SaaS
  3. SaaS-centric case in midmarket: main BI application in midsized businesses
  4. Elasticity case: BI for companies with strong variations in activity from season to season
  5. Power user flexibility case: BI workspaces are often considered necessary by power analysts

Then we analyzed the types of criteria which determine whether BI SaaS is a right or a wrong fit in your specific situation:

  1. The customization level of your data sources
  2. Where a BI project is on internal IT priority list
  3. Security, risk, regulatory, and privacy considerations
  4. Your application and organizational growth

Now on to vendor selection. You have two basic choices: your traditional BI vendors that are adopting their current product offerings to work in the cloud (not an easy task), or new specialized vendors offering BI products natively architected and built to operation in the cloud. The first sensible step for anyone seeking a BI SaaS solution is to check with their enterprise BI vendor on their SaaS offering. With a few exceptions, however, you may be quite disappointed. Most of the leading BI vendors are still playing mainly in the hosting (equivalent to Forrester SaaS maturitylevel 2) rather than in a true SaaS field. Specifically:

  • Early in 2010, SAP plans to lead the large vendor pack with self-service BI SaaS offerings, comparable to native BI SaaS products
  • IBM extends BI into private clouds based on its strength in app outsourcing
  • Oracle is mostly in the business of providing BI SaaS platform for ISVs
  • Microsoft is slowly but surely moving its BI platform into the cloud, piece by piece (SQL Azure)
  • SAS Institute is building its own cloud (in $70M, 38K sq foot hosting facility)
  • The rest of BI vendors concentrate on hosting domain specific BI SaaS solutions
  • Salesforce.com offers a respectable, though still-CRM centric BI SaaS solution

If these BI SaaS options from large vendors seem skimpy, your impression is correct. That's precisely why there's a plethora of small BI vendors addressing the need and the opportunity in the market left largely unaddressed by the mainstream BI vendors. The list is long and changes every few months based on newcomers and — yes, indeed — failures. Forrester compiled a long list that includes 1010data, Acsellerate Solutions, Analytix On Demand, Bime, Birst, Cloud9 Analytics, FusionOps, GoodData, Indicee, In2Clouds, Kognitio, myDIALS, Oco, Panorama Software, PivotLink, Rosslyn Analytics, SpatialKey, SRC, youcalc, and ZOHO.

Considering such a long and, as we mentioned, constantly changing list, how do you zero in on a winner? We offer a 3 step approach:

  1. First and foremost, it's all about business requirements. After all, even if the vendor has everything else right (e.g., reputation, stability, financial backing, strong management team, etc.) but does not support specific business requirements that you have, you're looking in the wrong place.
  2. Once a vendor passes business functionality check, we recommend bringing in an IT subject matter expert to help you understand the technical architecture under the covers
  3. The last, but not least, set of questions should be about the viability of the vendor. After all, most of these vendors are small startups, and some actually stopped operating this year

Please check out one of my recent blogs for the detailed criteria and the actual BI SaaS research report for detailed BI SaaS vendor scorecards.

Yes, the appeal of a BI SaaS application may be irresistible: You can subscribe and deploy it today, you don't need to go through the capital expenditure approval process, and you can even sometimes bypass cumbersome internal IT standards and protocols. Additionally, BI SaaS applications provide the instant BI workspace so desperately needed by power users and power analysts. But, unquestionably, there are risks, so we strongly recommend the following risk mitigation steps and strategies:

  • Perform due diligence for security, backup and disaster recovery
  • Do not overlook BI SaaS pricing and contract matters
  • Evaluate true long term cost of ownership
  • Double-check whether you may require additional source data licenses
  • Plan for the worst

I predict that BI SaaS adoption and market size will continue to grow for two main reasons. First, the lightning speed with which business, regulatory, and globalization requirements change can no longer be supported by traditional, often less-than-agile BI architectures. Forrester believes that BI SaaS coexistence case fits very well to support such requirement. Second, the availability of relational DBMS in public clouds will make it easier for vendors, IT architects, and application designers to move their own BI apps into the cloud.

Comments

re: BI In The Cloud? Yes, And On The Ground, Too

Great analysis, Boris. Very practical set of recommendations. To me, BI SaaS is where e-commerce was in the late 1990s: everyone thought the security risks were too great for it to ever catch on. Phooey then, phooey now! The only real obstacle is transferring large data volumes over a thin internet wire and unamortized investments in existing private data centers and premises-based applications.