Enterprise spend on BI tools to scale up

Increasing budgets for mature business intelligence (BI)
technologies (extraction, transformation and loading (ETL),
online analytical processing (OLAP)) may not herald an economic turn around, but
they are anindication that
practitioners see the need to integrate and analyze data as being strong,
perhaps stronger eventhan the
need to cut costs.

The Data Warehouse
Institute (TDWI) Giga Research Quarterly TechnologySurvey,
conducted at the TDWI Boston Conference on August 25, 2003, has surfaced
indications ofincreased optimism
in the business intelligence market. The data warehousing practitioners that
answered thesurvey anticipated
spending more for mature BI technologies ETL (62 percent) and OLAP (60 percent),
aswell as servers in support of
data warehousing (62 percent). More modest increases for the less mature BI
technologies data mining (34 percent)
and data profiling (38 percent) point to a continued sense of cautionwith
these useful, yet less proven emerging technologies.

There are other indicators hidden within these survey results,
as the largest segment of respondents remain undecided,
ranging from a low of 31percent for servers up to 63 percent for data mining.
However, there ismore optimism
in the numbers for increases greater than 5 percent, which are as high as 39
percent for ETL,40 percent for
OLAP and 44 percent for servers. In contrast, the expected budget decreases are
an order ofmagnitude lower for
all of these BI technologies, ranging from a low of 2.5 percent for OLAP to a
high ofonly 6.2 percent for
servers (see table below).

These results should not be seen as a sign of complete economic
turnaround in light of the fact that there is no mention
of new jobs being created. Rather, these results should be viewed merely as a
subtle shift in theperspective
of BI practitioners.

Recommendations

User organizations, and in particular IT, should increase
budgets for mature BI technologies, specifically data integration
and servers in support of BI initiatives. Approach emerging data mining
technologies cautiously,placing
an emphasis on predictive analysis solutions (fraud detection, customer churn)
that are targeted atbusiness and
casual users rather than tools that are targeted at the statisticians. Look to
data quality as a majortechnology
investment in the next 24 to 36 months, since bad data is more difficult to
integrate, wastes serverresources
and yields invalid historical and predictive analytic results.