The best-in-class may not be the best-for-you

Oracle User Group Survey: Achieving End-to-End Business Intelligence
November 15, 2006
Microsoft: supersize that data warehouse?
November 22, 2006
Show all

The best-in-class may not be the best-for-you

Onesheet_sm_1
In the IT industry there is always a lot of discussion (and
PR) regarding the best-in-class products. Ask any industry analyst, consultant
or IT person who keeping up with their reading and they can tell you the
best-in-class. These products are in research analysts’ upper quadrant and
always make everyone’s short-list. In business intelligence (BI), performance
management and data warehousing the best-in-class products are generally sold
by the largest software vendors (they may have built those products or they
acquired them but they have them now.)

That should make it very easy to determine what products to
evaluate and buy. And if you are in the Fortune 100 that is precisely what most
companies have done (although most companies have made the decision several
times so they have many of these products in their enterprise….that’s another
discussion.) However, a funny thing happened along the way, our best-in-class
vendors expanded the tool categories they develop (or acquire) and sell. Business
intelligence (BI) and data integration (formerly ETL or Extract, Transform and
Load) vendors sell suites. BI vendors sell reporting, OLAP (on-line analytical
processing), data mining, search and visualization tools in their suites. Add
in portals, metadata management, content management and maybe even ETL and the
BI suites are getting very robust. Do not forget a little professional services
and training for those suites too. It used to be you bought an ETL (Extract, Transform
and Load) tool but now you can get a data integration suite with ETL, EII
(Enterprise Information Integration), EAI (Enterprise Application Integration),
data quality, data profiling, metadata management tools available real-time or
batch. A pitch of professional services and training for those suites goes well
with that!

I think the BI and DI suites are terrific but has the
best-in-class been made into the best-for-some? Certainly many in the Fortune
500 can afford the software licenses, implementation costs and associated
hardware investments but how about everyone else, such as the SMB (small to
medium size market) or even many groups or projects in the large companies? And
it is not the cost of the software alone that makes these suites not palatable
for some, but the functionality built into these suites mean IT (or
consultants) along with the business users using the BI tools need to
understand a lot and take the time to get all the moving parts to work within
their environment. This means that these suites are too costly in terms of both
cost and resources to some.

If best-in-class is not best for you, what should you do?
Your choices include:

  1. Do nothing. It costs the least initially BUT business groups will fill the gap with data shadow systems! Business groups are running the enterprise and need to make decisions. If they cannot get the data or do the analyses they want they will built it themselves or make decisions based on their “gut”.
  2. Buy into the best-in-class vendors approach to the SMB market – pricing,
         product packaging, services packing (We’ll discuss these vendors’ approach in more detail at a later time.)a nd, more recently, some of these vendors have created SaaS (Software as a Service) offerings.
  3. Buy one of the many good to excellent BI or ETL tools offered by vendors other
         than on the “short-list”. There are many products that may be more affordable,
    from both a cost and resource perspective, which may enable a business
         solution with a great ROI. Many of these products have hundreds or even
         thousands of satisfied customers and are be sold by solid companies. Some of these companies may have even been around for a decade.  After all, many SMB companies are using enterprise applications from software vendors who are not on anyone’s  best-in-class list yet they work very well. Why should not BI, ETL, data  quality, data profiling or data modeling tools also work that way?
  4. A newer, but very interesting alternative is data warehouse or business
         intelligence appliances. This is a new wrinkle on the third option with some intriguing technology innovations.

    I am revising this post due to the excellent comments by Nicholas
    Goodman on BI
    . Based on his suggestion options 5 (SaaS) and 6 (open source
    BI) have been added as alternatives in addition to the original four that I
    posted. Thank you Nicholas.

  5. BI/DW SaaS (Software as a Service) offerings are starting to appear and are marketed as a very cost effective approach. It is important, however, to examine
    the architectural approach taken in these offerings and confirms that it will fit within your enterprise’s approach towards software and data. An example is whether your BI SaaS offering operates outside of your firewall while your data is within it? If so, will your enterprise buy into this approach? An intriguing variation on this theme is
    a combination of a SaaS offering provided on a DW/BI appliance within your firewall, with the SaaS provider remotely managing their appliance.
  6. Open source has become a viable alternative with many products in our industry
       such as are Linux, Apache, MySQL and FireFox. (fyi: DW/BI appliances are
         themselves often built upon several open source products, however, they
         usually add their own proprietary software to “juice” the offering.)
     

    Another revision based on Morgan Goeller’s insight
    (second comment). As per Morgan’s
    comments I am adding, with some qualifications, option 7. Thanks Morgan.         
                                                                                               
  7. Microsoft Excel!!! And maybe Microsoft SQL Server 2005 with Analysis Services thrown in? This will probably get me thrown out of the both the DW/BI industry
         analysts’ and architects’ clubs. I have written extensively about data shadow systems and have tried to do so in a pragmatic fashion with an attempt at understanding both the business and IT perspectives. I have worked with many clients in a wide variety of industries to “rehabilitate” their data shadow systems. Often, the final solution does indeed utilize Microsoft Excel as the “final” BI tool or presentation layer. However, I have also worked towards replacing (exorcizing) the data integration and data transformation aspects from the data shadow systems that were developed in a haphazard manner that are often accomplished with Microsoft Access, Microsoft Excel and SAS). We definitely need to discuss further since I do not want to be blamed for starting a new wave of data shadow systems.

Is the best-in-class the best-for-you? That’s you decision
based on what you can afford, as well as, what you value. After all, even the Consumer Reports
organization
lists short-lists for both a Best and a Best-Value category.

3 Comments

  1. I’d add “consider non traditional software offerings like Open Source and SaaS” as seperate from the SMB approach to the market.
    Perhaps you consider the open source alternatives part of the good to excellent a little lower.
    Regardless, it should be a consideration for those looking in the “best value” category.

  2. I would also consider adding the old standby of Excel/Access, especially as a BI Client. I don’t want shadow systems, not at all. Also, I am not a Microsoft fanboy (and am typing this on an iMac). So hear me out …
    I have seen shops that run with a low-cost, centralized database (say Postgres, MySQL, or even Sybase) and do analysis and reporting very effectively using Access/Excel. Very often, these products are already in an enterprise and can provide a great deal of flexibility combined with a reasonable amount of power for very little cost.
    It does require some thought and design (especially to consistency, data quality, and re-use) but often it can be done at low cost and with the resources and people on-hand. This is critical for the small/medium business who is just starting to see the need for BI and doesn’t yet have the budget for it.
    The reason why these tools end up being used for shadow systems is precisely because they are easy to get started with. Why not co-opt them into the system before you get started?
    Worth considering, anyway.

  3. Oracle Plunges and no one soars in the 2006 Gartner Data Integration Tools Magic Quadrant

    Gartner has turned the Magic Quadrant for ETL tools into Data Integration Tools and in the 2006 report Oracle has dropped from a leader to a niche player.

Leave a Reply to An Expert's Guide to WebSphere Information Integration Cancel reply

Your email address will not be published.