Populating the model should be approach as any other data
migration exercise - for that is exactly what it is.
Phase 1- Scoping
This phase should be a very short initial piece of work to
provide a first level assessment of the scope and cost of the data migration.
It is where the source data is analysed, the scope of the
work is determined and the overall risk is evaluated in terms of the complexity
of the work involved.
The primary deliverables from the Scoping phase are:
A high-level estimate of the number of days and cost to complete
the data migration.
A high-level statement of any major issues identified at this
stage that may affect the delivery of the migration.
A high-level work plan for the migration and estimated
An estimate of the resources required to support the migration.
Provides a specific estimate of the number of days and cost for
the subsequent Discovery phase.
Provides a budgetary anticipated final cost estimate for
Phase 2 - Discovery
This phase comprises the detailed work necessary to
determine the exact costs and plan for doing the migration exercise. It ensures
the Enterprise understands the scope of what is being undertaken, including
detailed plans and costing’s.
The primary deliverables from the Discovery phase are:
Agreement of the Meta-model to be populated.
A detailed understanding of the scope of the migration activity
that is to be undertaken, including identification of the Datasources (source
files, locations, formats, etc that are being migrated from).
An understanding of the quality of the existing information and
the options available for data cleansing.
Identification of any hardware and software that will be
Documentation of the issues and risks with the proposed
Identification of the key resources from each of the parties
involved, who will be required to ensure success, including an assessment of
resource capability and the training that may be required.
Production of realistic plans and costs for the work.
Phase 3 - Provision
This phase comprises the work required to gather together
all of the information that currently exists into one place in preparation for
the analysis phase.
It is likely that as this phase progresses other information
sources previously unknown will be discovered.
Phase 4 - Analysis
This phase is the core of the work that needs to be
undertaken which invokes analysing and understanding the information gathered.
This analysis will consider the following:
Format - The format for fields
may not be consistent. E.g. a field meant to indicate the cost of maintenance
may be expected to contain values in thousands (10), but may contain the actual
Structure - There may be
structured data in free-format fields which therefore contain information that
should be structured but are not.
Duplication - This is the most
common area where problems can exist. There may be two, three, or even more
“lists” of the same information. E.g. more than one group has a list of
Validity - The values that are
contained in fields may or may not confirm to what was originally meant. E.g. a
field meant to indicate the importance of an application to the business may be
expected to contain values such as “C” for Critical, “I” for Important, “N” for
No important, but may contain unexpected values such as “X” or “-“
Currency - The information may
be recent or may be out of date.
Integrity - There may be related
information that should exist but does not. E.g. An application may say it
utilises a database but the database may not be defined.
Integration - Will the
information in the tool replace the information in its current form or require
integration and synchronisation. (To be carried out in Stage 4 of the Modelling
Phase 5 - Cleansing
Having analysed the information in the previous phase, this
phase takes this knowledge and uses it to manipulate and process the information
into valid, current and consistent data suitable for importing in the Model.
This is where difficult decisions regarding which
information is “correct” or not are made and therefore the time and resources
required to perform this critical phase should not be underestimated.
If this phase is not performed or is done badly it makes the
entire exercise utterly pointless. Most Enterprises ignore this phase which may
go somewhere to explain why many modelling efforts fail.
Phase 6- Loading
This phase is concerned with loading the information into
the tool’s repository and the associated checks and testing to ensure
correctness completeness and consistency.
Does your Enterprise approach
modelling as a Data Migration exercise?
Are there clear responsibilities
with respect to Scoping, Discovery, Provision, Analysis, Cleansing and
If you do not do one or more
of the above, does that cause issues and problems?
If so, what are the effects of
those issues and problems?
What will you do to solve