|
The third wave of Enterprise data integration is beginning.
I. The first wave was placing operational data into warehouses in hopes it would
be used for insight.
II. The second wave was extracting this data into data marts where dimensional insights
could actually be generated.
III. The third wave is to put the computed insightful metrics back into the warehouse
and onto the corporate bus for use across the corporation.
There is a culture gap between warehouse technologies and data mart technologies
that inherently inhibits sharing “marthouse” (my term) data back into warehouses
– see below “Why is mart and warehouse integration not being done?” The technologies
exist, but what this era needs is use cases and best practices that bridge the culture
gap and provide the New Knowledge that cannot be gained in only the mart or warehouse
domains alone.
This will turn the culture gap into a spark gap which ignites new competitive insights
heretofore unavailable.
Examples and use cases are needed to ignite this third wave of integration.
Following is a transportation industry example of the untapped power inherent in
merging specific data mart and warehouse data – at the model level:
You are an airline executive. Your majority stockholder greets you for lunch with
the question, “So, did we make money on my flight here from Chicago this morning?”
There is a long silence...
Let us look at how the data you already have can answer this question by unifying
currently separate architectures.
- A multidimensional data mart such as Essbase serves uniquely by allocating expenses
down to each flight.
- But other factors of profitability change on an hourly basis. Data on headwinds,
fuel prices at each departure airport, gross flight revenue based on actual boardings
and actual fares, which flight crew flies, and so on are instantaneously variable.
- These metrics are obviously captured in the warehouse or the ODS.
By merging allocated planning expenses from an Essbase model by flight with the
instantly changing operational warehouse data, it is possible to know actual profitability
by flight in near real time.
One can then report by factors it would be very unlikely to put into an Essbase
planning model, such as Pilot Name (may change at the last minute), headwinds, instantaneous
fuel costs in the airport of departure, and so on. You can see in near real time
what routes are profitable, which flight crews save you money, where air traffic
control delays affect profitability, etc. And you can see the direct and indirect
components of profitability as they occur.
You need Essbase for allocating the costs and assumptions, direct expenses, and
indirect expenses such as marketing expenses, all corporate overhead. You need the
warehouse for the instantaneous operational data.
By merging data mart members and data warehouse data columns at the model level,
new models can be developed. New knowledge is revealed which could not have been
derived without both multidimensional and table data.
The result delivers up to the minute operational profitability. “Real time” adjustments
can now be made:
- Refuel upstream, in airports where fuel is less expensive.
- Use this option to renegotiate better fuel costs in expensive departure cities
- Require corrective action from station managers where flights lose money due to
late departures, excessive runway idle time, etc. • Counsel flight crews on money
losing tasks.
- Etc.
Every parameter reported becomes an opportunity for tuning profitability “on the
fly”.
|