continued...page 4 of 4 Clearly MS/SSOS and the new Microsoft standards (see story, "Decision Support Cycle Unique") have to concern the MOLAP and
HOLAP vendors. Though MS/SSOS is still in beta testing, I believe its performance will be at
least competitive to incumbent MOLAP/HOLAP tools, even though it may lack some of their
functional nuances. MS/SSOS's aggregation-tuning wizard and partitioning capabilities certainly
are innovative.
Because the multidimensional OLAP design is constantly shifting and changing, tuning an OLAP
database can be part art, part science. The Microsoft wizard puts a bit more science into the art.
And as Netscape has learned, ever-improving Microsoft freeware does win market
share.
There are some natural limits to how much business Microsoft will steal from MOLAP/HOLAP
vendors. Developed applications can be vital OLAP components--and MS/SSOS has very few of
those. The mechanism for connecting MS/SSOS to data sources is OLE/DB for OLAP, and it
requires custom programming. Other vendors' tools provide non-programmer power users with
simpler scripting languages. In addition, MS/SSOS runs only within Windows NT, and that
operating system's scalability, reliability, and performance usually don't match Unix and other
mainframe systems. There's no built-in Web interface; third-party tools will have to supply one.
MS/SSOS will not be supported directly by Microsoft with a set of consultants who can provide
vital support and downstream services. Also, beta 3 of SQL Server 7.0 doesn't have cell
writeback and cell security critical to budgeting and financial modeling and what-if analysis;
the shipping product may or may not have it.
Clearly, MOLAP/HOLAP vendors will have fewer niches than their data warehouse-building
counterparts. And given that the OLAP client and desktop vendors have had the OLE/DB for OLAP
interface to work with for nearly a year, expect some amazing offerings in the new OLAP front
ends to MS/SSOS. Already there's the free Seagate Workshop. It's important to note these new
front ends will also work with the SAS MDDB engine, Whitelight Server, and any other OLAP
servers that become OLE/DB for OLAP providers as well as consumers.
While end users may be happy with the pricing windfalls and new interoperability, OLAP vendors
may have to find a new way to find profit. However, this is an idiosyncratic market with a lot of
specialized niches where the possibility of creating added value is possible.
Important things are happening in P&DSS. Data mining and knowledge discovery are not only
feeding better information and insights into planning and marketing strategy. They're also acting
as support services for industry-specific operational processes. Likewise, very large data
warehousing systems are becoming critical targeted-marketing planning systems for many
national retailers. With Web implementations or E-mail--and pager-based systems such as
MicroStrategy's Broadcaster--organizations can trigger highly specific reports and alerts to
hundreds of key people based on weekly or even daily updates.
With key need-to-know triggers and alerts filtering throughout an organization, P&DSS systems
are starting to map back closely to operational systems by providing informed feedback that
makes for high payback opportunities. Likewise in OLAP modeling, more vendors are going beyond
raw software with turnkey applications such as Gentia's Balanced Scorecard or Hyperion's
OnTrack.
Microsoft's SQL Server 7's free P&DSS will change the industry. The server's release also raises
the antitrust question: Is it fair to be able to bundle and price at zero in markets where a vendor
has little or no presence (OLAP) while raising prices in markets in which it has a monopoly
(desktop operating systems)?
The free services are also a test of whether the classic PC marketing model--which is to set
prices low in large, growing markets while making the software easy to use and offering some
free, competitive capabilities--can win significant P&DSS share in a market where IS shops are
used to round-the-clock availability, scalability, and maintainability--plus close consulting
support. Finally, will software vendors with distinctive specialties in data warehouse building
or niche OLAP competencies be able to thrive and prosper when faced with free features from a
major vendor?
Jacques Surveyer is a consultant based in Toronto. He can be reached at jbsurv@inforamp.net.