Mike Gadbury, Vice President Aremissoft and Velibor Korolija, Operations Director, Bromley Group Research Organisation
What is Data Warehousing?
Data warehousing means taking information from PMS, budgets, spreadsheets, marketing
information, etc. and copying it into a separate central warehouse where it is
reorganised, cleaned and re-loaded onto another database that is tailor-made for
exploitation.
What is Data mining?
It means data is scrutinized by financial and marketing analysts to detect relationships,
trends, customers segments, etc. which can be used for -
w Customer Relationship Management eg. how effective was
a media advert, who are the 100 best customers, how many customers come back etc.
w Enterprise performance measurement activity based
costing, benchmarking, performance indicators, strategic planning.
w Enterprise resource management eg. greater efficiency
with suppliers, relating the cost of labour to level of service, inflation etc.
Recent progress on data warehousing
w Data warehousing is consolidating rapidly and mature
reliable products are now replacing first generation tools.
w It is moving towards a situation where you buy 80%
features and build less than 20% (you would only build if you had complex, incompatible
back end systems).
w It is becoming more heterogeneous - multiple products and
decentralised buying decisions favour this.
w Closed-loop systems are developing, which allow decisions
using data warehousing information to be fed back into transactional applications or the
data warehouse itself (eg campaign tracking applications).
w E-commerce is helping to collect 'clean' information
because customers enter data themselves. It allows for 'clickstream analysis' of the
customer's progress through the site. (The Microsoft site generates 600 gigabytes of
clickstream info every day!)
w Beware of vendors who claim to have developed an
information portal, ie. a centralised, single, uniform access to all corporate
information. Be sure when buying the pieces in your corporate jigsaw that they comply to
open standards, to web protocols, XML whatever is appropriate.
w Microsoft has entered the market with Sequel Server 7
which provides certain services on which other vendors can build applications eg. data
warehousing.
Questions to ask when building your data warehouse
w Buy or build? Build only when you need features that are
not available otherwise
w When you build, beware of time and cost implications
w Use internal IT dept. or outsource? Look at the skills of
your own dept outsourcing will probably be more logical
w Is your data warehousing strategy aligned with your
business strategy? Ensure the buy-in of senior management
w ROI what strategic advantage will it bring? This will
be difficult to assess but you can assess the savings of consolidating and distributing
information
w How to begin? Start with a small pilot information from
a couple of internal systems eg. marketing or finance. When mastered, you build up
w Single vendor or 'best of breed'? Look at your current
strategy. If you already work with Oracle, go with them but beware of vendor dependence.
If you chose a best of breed solution, make sure the pieces are compatible both now and in
the future
w Which tools to use? Organise user workshops. Ensure all
business people understand your corporate goals. Organise a series of vendor demos. Be
aware of their long-term viability and ask 'is this vendor going to be around in 5 years
time?' Fight against jargon make vendors comply to templates
w Be aware of the potential benefits but also the risks
Vos commentaires : cliquez sur le Forum des Blogs des Experts
L'HÔTELLERIE Eurhotec Special Issue 8 February 2001