Alacrity Results Management ARM

1
Alacrity Results Management (ARM)

• Alistair Davidson
• Eclicktick Corporation www.eclicktick.com
• Howard Weinstein
• Cherniak Software www.cherniaksoftware.com

Draft 3.1 November 5th, 2003
2
Alacrity Results Management

• ARM innovative proven support for performance
  improvement with superior value and capabilities
• Lower cost
• More capabilities
• Higher extensibility
• Distributable over multiple servers
• Possibility of software agents and genetic
  modeling attached to system

3
What is Alacrity Results Management Version 7.303?

• A component based object oriented end userized
  application that provides
• Information warehouse collection from
  heterogeneous data sources
• Large scalability and robustness
• Modeling capabilities that can managed complexity
• Consolidation capabilities
• Modeling with different models working off same
  granular data
• Performance reporting and scorecards
• Integration with Excel for reporting
• Rapid construction and modification
• Batch construction via descriptive meta-language
• Manual construction

4
Why Is ARM Different?

• Developed by innovators in the strategic planning
  and strategic IT planning space
• Has been used for large scale data collection,
  budgeting, consolidation, reengineering data
  tracking, IT portfolio tracking and market
  modeling
• Designed to deliver fast large and complex
  projects on impossible deadlines
• Designed to eliminate the time, costs, checking
  and frustrations of trying to make multiple
  technologies work together
• Designed to be a long term sustainable
  architecture for rapidly changing problems
• The underlying database technology is in
  production with 200 large companies
• ARM is in production in large multi-location
  environments where it is simultaneously being
  used as an information warehouse for low level
  data and also for modeling, performance tracking,
  budgeting and consolidation.

5
Why ARM?

• Less expensive
• 10-20 of equivalent solutions
• More flexible
• Supports multiple views of data
• Quicker configuration
• Designed to support rapidly changing needs and
  impossible deadlines
• Batch construction by a financial analyst
• Scales better
• Federated
• Different paradigm than OLAP or relational
  databases
• Allows storage of raw operating data without
  performance penalty minimizing impedance
  mismatches between RDBs, OLAP datacubes and
  object architectures
• Modern architecture
• OO principles applied to business intelligence
  and business activity monitoring
• Permits many to many relationships unlike
  relational databases
• Does not experience exponential growth problems

6
Why Is It Less Expensive?

• One solution provides
• Integration with spreadsheets for input and
  output
• A central information repository that can combine
  data normally kept in different places
• An architecture that does not bog down as size
  grows
• Essentially, you get an information warehouse for
  free

7
ARM Uses a Node and Link or Network Architecture

• Nodes are smart objects that can be thought of as
  small OLAP cubes, preprogrammed to deal with
  time, accounts and variants, rather like a 3-D
  Excel spreadsheet with more functionality
• Nodes can be linked in any way you want
• You can have multiple and inconsistent linking to
  represent the complexity of organizations
• e.g. linking organizational structure with
  processes with products with markets with projects

8
Smart Numerical Objects That Know
Numerical item
Versions
Formula management and formula cloning
Repository data
Time series roll up
Consolidation and modeling formulae
9
ARMs Smart Object Structure

• A simple hierarchy
• Plans consist of
• Statements consist of
• Items have
• Item numerical calculations and
• Meta-information
• Items are associated with one or more parent nodes

10
ARM Has Fewer Limitations Than Commonly Used
Technologies

• Any type of item
• Any type of chart of accounts
• Any number of consolidation paths or owners of a
  numerical item
• Any number of business units
• Any number of consolidation layers
• No actual software size restrictions other than a
  billion objects per server
• In the interests of full disclosure, an ARM smart
  object consists of a group of objects, but
  databases can be spread over multiple servers to
  get around the billion objects per server
  limitation

11
Example Why Is This Important for Client X?

• Significant development time in Hyperion (cost
  around 250K of programmer time over 3 years)
• Current model probably close to limit to
  practicality
• Future growth problematic
• Processing takes 10-12 hours (down from 24 hours)
• Five cubes need to be maintained
• A 10 gig. eSSBase model demonstrates the problem
  of a 9 dimensional representation of the business
• The ideal system mapped out by planning would
  probably not be implementable directly in
  Hyperion as it requires more integration of
  information than is current achievable
• The model is so large that only 10 of the actual
  data is stored. 90 has to be calculated for size
  reasons.

12
Client X

• The really expensive part of the process is the
  roughly 30-45 person years of time for
  maintaining the planning system every year
• The even larger expense is the unmeasured cost of
  missed opportunities from inability to tie
  together
• trends and costs,
• markets and opportunities,
• promotions and channel activity,
• production and procurement

13
An object approach allows more dimensions than an
OLAP tool
Organizational unit
Customer object
Scorecard object
Employee object
Product object
Skills knowledge attributes
Date object
Statement object
Process object
Market research object
Financial object
Contract object
Project object
Numerical object
Source object
Resource object
14
Multiple Consolidation and Modeling Trees off
Same Source or Aggregated Data
Shadow consolidation
Business unit
Project
Business unit
Channel
Process
Mfg. data
Customer segment
15
ARMs Network Model Architecture

• Does not grow exponentially like an OLAP cube
• Easily handles many-to-many relationships that
  are impossible in relational databases (RDBs)
• Allows raw data to be used by multiple
  applications and views
• Once consolidations and aggregations are
  calculated they dont need to be recalculated so
  raw data can be mixed with aggregated data
• Control over incremental and batch processing
  also an option
• Virtual database architecture provides
  flexibility in processing power
• Will be increasingly important with grid computing

16
ARM Scales Robustly

• ARM is based upon industrial strength database
  technology that has been around for over 20 years
  (Gemstone Object Database)
• ARM has smart recalculation so large size does
  not mean long recalculations
• Unlike OLAP or RDBs, ARMs processing does not
  slow down exponentially as complexity increases
• ARM has superb auditability and flexible drill
  down so the source of information is always
  visible
• ARM can be constructed in batch using a
  metalanguage in Excel. This permits use of
  standard Excel functionality for constructing
  repetitive knowledge models and instances of
  smart objects
• Surprisingly, existing clients with ARM dont
  need DBAs.

17
We dont like to overclaim

• For a number of years, Gemstone has supported a
  virtual database capability so data can be spread
  across networks
• Local countries can have their own repositories
  and models
• Head office can grab data without caring where it
  is located
• ARM requires only trivial modifications to take
  advantage of this feature in Gemstone

18
Modern Architecture

• Object oriented design is critical to controlling
  software costs
• Spreadsheets are not typically OO in design
• Neither are RDBs
• ARM delivers the benefits of OO and has an order
  of magnitude lower cost of ownership

19
Sources of cost advantage

• Elegance
• One solution can provide
• An information warehouse
• Consolidation
• Modeling
• Cost allocation
• Reporting
• Integration with spreadsheets
• A repository for flat files from multiple data
  sources
• An extensible architecture
• Agents, web front ends, integration with
  relational databases are all doable at low cost

20
ARM Savings

• Reduced testing time
• Reduced number of tools
• Reduced maintenance costs
• Faster development times
• Elimination of DBA function
• Less expensive sourcing of data for new
  applications and reports
• Reuse

21
The Big Benefits (1)

• A better view on the problem leading to
• Lower costs
• Better modeling
• Lower cost of changing analysis
• The ability to analyze what you need to analyze
  when you need to do it
• Massive integration abilities at low cost
• An architecture that can scale globally
• Built in business activity monitoring (BAM)
  functionality

22
The Big Benefits (2)

• Performance report (red, amber green) for
  scorecards
• Superb drill down that matches or exceeds OLAP
  capabilites
• High integration with existing spreadsheets (in
  and out)
• Faster calculation with more data in the system
• Lower maintenance costs

23
Example Modeling a Telecom Fixed Wireless
Vendors National Market by City by Customer by
Service Down to Capacity and Equipment
Consumption, Costs and Profitability
Residential demographics
Penetration curve
Bandwidth requirement
Customer premise equipment
Business customers
Sales activities
Customer service activities
Node deployment
Customer installation
Cost of acquisition
Marketing costs
Network capital costs
Customer profitability
Profitability by customer, node, network or
business unit
24
The underlying technology in ARM

• Gemstone
• 200 major customers on 6 continents
• State of the art object management with
  transaction processing options
• Over 20 years of leading edge development
• Alacrity Results Management
• Cited by the Convergent Engineering Institute as
  one of the Top 7 object applications in the world
  as an example of deployed object representations
  of enterprises

25
Consider the following

• What kinds of opportunities are your missing
  because
• You cant tie together informational silos
• Market research to sales to promotions to
  inventory to production to marketing to finance
• It takes too long to do certain kinds of analysis
• You spend most of your time checking data rather
  than making decisions based on data
• Too many people are required to find an answer
• Too much time is spent on maintaining value
  reducing mappings e.g.
• RDB OLAP Spreadsheet OO Model

26
Contact Information

• Alistair Davidson, Eclicktick Corporation
• 29 Clinton Street, Suite 305
• Redwood City, CA 94062
• Phone [phone]
• alistair_at_eclicktick.com
• Web site www.eclicktick.com
• Howard Weinstein, Cherniak Software
• 10 Commerce Valley Dr. East, Suite 400
• Thornhill, ON L3T 7N7, Canada
• Phone [phone]
• howard_at_CherniakSoftware.com
• Web site www.cherniaksoftware.com