# Proposal, Cost Estimate, Consulting by quu17210

VIEWS: 125 PAGES: 126

• pg 1
```									     ITPM Today—3/2/2010
1.PERT REVIEW (last part of Ch
7)
2. The Proposal—due Friday.
10/10
3. Time and Cost Estimation (Ch.
8, copy packet)
PERT (Program Evaluation and
Review Technique)
 Not supported by MS Project
 Gives you probabilities of completion of a
project by a certain time
 Calculates a standard normal random
variate and uses a probability table to find
its probability
PERT INPUTS
 A = most optimistic time
 M = most likely time
 B = most pessimistic time
PERT formulas
 Task mean = (A + 4*M + B)/6
 Task Std. Dev. = (B-A)/6
 Project mean = sum of all the task means of
 Project std. Dev = sum of all the task
standard deviations of tasks on the critical
path
PERT Homework Assignment
THE PROPOSAL
 You want it by When????
 Primarily a selling document
 I will grade it on the basis of:
– How well you sold the product
– How well you sold yourself
 Are cost and duration the two most
important parameters in the Proposal?
 Use the format outlined in Chapter 11!
Proposal
    Address it to the customer, not me (your
instructor)
solution
2. Describe why you or your firm should be the
one chosen to do this excellent work
Proposal Methodology in the real
world
 Your proposal will be reviewed and
approved in-house before it is allowed to go
to the customer
 Sometimes a proposal manager conducts the
review
 Budget gets checked
 Boilerplate must be there to protect the
proposer/contractor
PROPOSAL CONTENTS
A. COVER LETTER                             (SEPARATE PAGE)

B. TITLE PAGE                      (SEPARATE PAGE)

1. Scope            (SEPARATE PAGE—SCOPE OVERVIEW, FOLLOWED BY YOUR PROPOSED SOLUTION)

2. Advantages                   (OF YOUR PROPOSED SOLUTION AND WHY YOU SHOULD BE SELECTED
TO DO THE WORK)

3.    Financial          (USE SAME BREAK-EVEN CHART YOU USED IN YOUR PROJECT PLAN)

4.    Plan        (JUST AN OVERVIEW HERE)

5.    Deliverables                (AND THEIR DUE DATES)

6.    Acceptance                (DISCUSS YOUR ACCEPTANCE METHODOLOGY HERE)

7.    Alternatives
8.    Terms, Conditions and Assumptions
9.    Terminology
Cover Letter
Follow the style exhibited in the copy
packet, chapter 11
Estimation: An ART
McDonnell/Douglas Aircraft) wealthy
 Accurate intelligence information is a help
INPUTS             TOOLS                 OUTPUTS

1. Activity list   1. Expert
2. Constraints
Activity
1.
judgment
3. Assumptions                          duration
4. Resource        2. Historical data
requirements       3. Analogous         estimates
5. Resource        estimating
capabilities       4. Simulation
6. Historical                           2. Basis of estimates
information
The Cost estimation Story—
Steve McConnell
   You can’t tell exactly what its going to cost

until you know exactly what   IT is.
   (Which is why we spent so much time
conceptualization and definition)
A Rule of Thumb
 The time to design, document and code a
module =
 equals the time to debug it (TEST IT)

   According to Gildersleeve
Estimating Rules (Rakos)
   Never use inexperienced persons to estimate
   Get group estimates if possible
   Never force an estimate on a programmer
   Never take an average of different estimates
   Granularize down to FOUR or less weeks, roughly
   Always quote a range when giving estimates
Rakos’ Conclusions to
Estimating
 Our weakest talent
 Estimating is iterative
 Estimating is still an art
Review: Project Time
Management Processes
   Project time management involves the processes
required to ensure timely completion of a project.
Processes include:
– Activity definition
– Activity sequencing
– Activity duration estimating
– Schedule development
– Schedule control
   Which of these gets performed in the Planning
and Budgeting Stage??
   Assign the task to a project player
   Ask the player how long it will take him or her to
complete the task (This gives the player ownership
in the planning)
   Player provides their best estimate
   The player understands that they will be required
to complete the task within their estimate—their
feet will be ―held to the fire‖
Time Estimation: Goldratt
 Claims team players add safety to their
estimates
 What is safety??
 Can determine how much by asking the
question, ―How sure are you that you will
Time Estimation--programmers
   Naïve programmers have a horrible
durations and costs
Time Estimation—making time
for creativity: SLACK
 Keep in mind that customers unintentionally
put projects under extraordinary schedule
pressure—more for less
 A consideration in schedule development is
to take the tasks requiring creativity and
place them on ____?!?
Too little time syndrome

More stress

More mistakes
More schedule pressure

Needed: A Rule-based Expert
   IF ESTIMATER IS SEASONED AND IF THE
WORK PACKAGE REQUIRES CREATIVITY
ON THE PART OF THE ESTIMATOR, THEN
LEAVE ESTIMATE AS IS.
   IF ESTIMATER IS NOT SEASONED AND
ESTIMATE APPEARS TO BE OPTIMISTIC,
THEN INCREASE ESTIMATE BY 30%.
ANOTHER AI RULE
   IF ESTIMATOR IS SEASONED AND
ESTIMATOR ASSERTS 90% OR ABOVE
CONFIDENCE HE WILL COMPLETE
WORK WITHIN HIS ESTIMATE AND IF
WORK PACKAGE DOES NOT REQUIRE
SIGNIFICANT CREATIVITY, REDUCE
ESTIMATE BY 50%
Time Estimation
   What are the three approaches to time
estimation??
– Expert judgment
– History database
– Computer model or formula
Lefkon’s Methodology
1.   Divide the software project into as many individual
2. Predict the level of effort required to complete each task and
multiply that prediction by 2.
3. Add up the numbers and multiply by 2.0 again to account for
testing and debugging.
4. Take the total and multiply by 1.25 to account for meetings,
5. Multiply this level of effort by your company’s ―magic number‖
for labor costs.
Lefkon’s Methodology
6. Present this to management as a range. Take the
cost as predicted above and present the range as –
10 percent and +25 percent.
7. Stand your ground and remind management that you
did not arbitrarily come up with these numbers and
they cannot be adjusted arbitrarily. You may have to
suggest reducing scope and cost if management
does not agree with your estimate.
8. Revise your project budget as you undertake and
complete the project.
Project Cost Management Processes
   Resource planning: determining what resources and
quantities of them should be used
   Cost estimating: developing an estimate of
the costs and resources needed to complete
a project
   Cost budgeting: allocating the overall cost estimate
to individual work items to establish a baseline for
measuring performance
   Cost control: controlling changes to the project
budget
Cost Estimating
We need to speak the language and
understand the terminology:
ROI, IRR, NPV, Sunk costs
Tangible and intangible costs
Direct and indirect costs
Learning curve theory
Reserves (\$ included in a cost estimate to
mitigate cost risk; also called contingency
reserves)
Cost Estimating
 An important output of project cost
management is a cost estimate
 There are several types of cost estimates
and tools and techniques to help create them
 It is also important to develop a cost
management plan that describes how cost
variances will be managed on the project
Proposal Pricing Strategies--
Kerzner
   Type I Acquisition: One of a kind project
with little or no follow-on opportunity
» win the project, perform well and make a profit
   Type II Acquisition: New Program with
potential for large follow-on business or
representing a desired surge into a new
market
» win the project, perform well and gain a foothold in
a new market segment, usually at a loss
Table 6-2. Types of Cost
Estimates
Type of Estimate        When Done                 Why Done          How Accurate
Rough Order of       Very early in the         Provides rough         –25%, +75%
Magnitude (ROM)      project life cycle,       ballpark of cost for
WAG                  often 3–5 years           selection decisions
SWAG                 before project
completion
Budgetary            Early, 1–2 years out      Puts dollars in the    –10%, +25%
budget plans
Definitive           Later in the project, <   Provides details for   –5%, +10%
1 year out                purchases, estimate
actual costs
Estimation in General—COST or
TIME
 History data base
 Expert judgement
 a model like COCOMO
Cost Estimation Tools and Techniques
   4 basic tools and techniques for cost estimates
(Schwalbe—Ch 6):
– Analogous or top-down: use the actual cost of a previous,
similar project as the basis for the new estimate
– Bottom-up: estimate individual work items and sum them
to get a total estimate
– Parametric: use project characteristics in a mathematical
model to estimate costs
– Computerized tools: use spreadsheets, project
management software, or other software to help estimate
costs
Constructive Cost Model
(COCOMO)
 Barry Boehm helped develop the COCOMO
models for estimating software development
costs
 Parameters include source lines of code or
function points
 COCOMO II is a computerized model
available on the web
 Boehm suggests that only parametric models
do not suffer from the limits of human
decision-making
Typical Problems with IT Cost
Estimates
   Developing an estimate for a large software project is a
complex task requiring a significant amount of effort.
Remember that estimates are done at various stages of the
project
   Many people doing estimates have little experience doing
them. Try to provide training and mentoring.
   IT People have a bias toward underestimation. Review
estimates and ask important questions to make sure
estimates are not biased
   Management wants a number for a bid, not a real estimate.
Project managers must negotiate with project sponsors to
create realistic cost estimates
Category
Project Cost Estimate Overview
Description
Objective                         Install a suite of packaged financial applications
software which will enable more timely
information for management decision-making,
allow for cost savings through productivity
improvements throughout the company.
Scope                             The core financial systems will be replaced by
Oracle financial applications. These systems
include:
 General Ledger
 Fixed Assets
 Ops Report [AU: spell out Ops]
 Accounts Payable
 Accounts Receivable
 Project Accounting
 Project Management
Assumptions                       Oracle's software provides

   Minimal customization
   No change in procurement systems during
accounts payable implementation
Cost/Benefit Analysis             BSR was broken down into a three-year cash
& Internal Rate of Return (IRR)   outlay without depreciation. Costs are
represented in thousands. Capital and expenses
are combined in this example.
Project Cash Flow Analysis
FY95       FY96       FY97         3 Year    Future Annual
Total    Costs/Savings
(\$000)     (\$000)     (\$000)       (\$000)        (\$000)
Costs
Oracle/PM Software             992        500            0      1492               0
(List Price)
60% Discount             (595)                             (595)
Oracle Credits           (397)          0                  (397)
Net Cash for Software            0        500                    500
Software Maintenance             0         90        250         340             250
Hardware & Maintenance           0        270        270         540             270
Consulting &Training           205        320          0         525               0
Tax & Acquisition                0        150         80         230              50
Total Purchased Costs          205       1330        600        2135             570
Information Services &         500       1850       1200        3550               0
Technology (IS&T)
Finance/Other Staff            200        990        580        1770
Total Costs                    905       4170       2380        7455             570

Savings
Mainframe                               (101)      (483)       (584)            (597)
Finance/Asset/PM                        (160)     (1160)      (1320)           (2320)
IS&T Support/Data Entry                  (88)      (384)       (472)            (800)
Interest                                    0       (25)        (25)            (103)
Total Savings                           (349)     (2052)      (2401)           (3820)

Net Cost (Savings)             905       3821        328        5054           (3250)

8 Year Internal               35%
Rate of Return
Cost Budgeting
 Cost budget involves allocating the project
cost estimate to individual work items and
providing a cost baseline
 For example, in the Business Systems
Replacement project, there was a total
purchased costs estimate for FY97 of
\$600,000 and another \$1.2 million for
Information Services and Technology
 These amounts were allocated to appropriate
budgets as shown in Table 6-5
Table 6-5. Business Systems Replacement Project
Budget Estimates for FY97 and Explanations
Budget Category      Estimated Costs                    Explanation
Headcount (FTE)                     13   Included are 9 programmer/analysts, 2
database analysts, 2 infrastructure
technicians.
Compensation                \$1,008,500   Calculated by employee change notices
(ECNs) and assumed a 4% pay increase in
June. Overload support was planned at
\$10,000.
Consultant/Purchased          \$424,500   Expected consulting needs in support of the
implementation efforts; maintenance
expenses associated with the Hewlett-
Packard (HP) computing platforms;
maintenance expenses associated with the
software purchased in support of the BSR
project.
Travel                         \$25,000   Incidental travel expenses incurred in
support of the BSR project, most associated
with attendance of user conferences and
off-site training.
Depreciation                   \$91,000   Included is the per head share of
platform depreciation, and the depreciation
expense associated with capitalized
software purchases.
Rents/Leases                   \$98,000   Expenses associated with the Mach1
computing platforms.
Other Supplies                \$153,000   Incidental expenses associated with things
and Expenses                             such as training, reward and recognition,
long distance phone charges, miscellaneous
office supplies.
Total Costs                 \$1,800,000
Designing the Baseline
   One of the most crucial inputs to the pricing
decision
   Baseline design should be started early so its cost
estimates can be included in the proposal
   Effective pricing should begin a long time before
proposal development
– Gives management an opportunity to terminate a bid
initiative before too many resources get committed to
proposal development, presentations, negotiations, etc..
Pricing Process
 This activity schedules the development of
the work breakdown structure and provides
management with two of the three
operational tools necessary for the control
of a system or project
 The third tool is the WBS
The WBS as price estimating tool
 Provides the basis for effective and open
communication between functional
management and program/project
management
 After pricing is complete the WBS forms
the basis of a communications tool by
documenting the performance agreed on in
the pricing effort.
Organizational Input
Requirements
   After the WBS and activity schedules are
established, an organizational meeting is called.
– The WBS is described in depth
– Responsibilities are clarified
– Costing information is solicited and collected from the
responsible parties
   A short time fuse is usually involved in
estimating/pricing which makes it all the more
risky
– RFP’s sometimes require a response within 30 days of
their submittal
Labor Distributions
   Functional units supply their input in the form of
man-hours
» See Figure 14-2
   Man-hours submitted are often over-estimated
   Man-hours are converted to dollars by multiplying
by the labor rates
   Rates are only averages
   Base rates are then escalated as a % factor, based
on past experience
Labor Distributions--Conflict
Resolution
 The reduction of man-hours is often the
source of heated discussions between
project and functional management
 Most common solution rests with the
project or program manager
 This becomes the usual turf-fight
 How would you resolve all such
conflicts???
A Proposal Manager
 Integrates the activities of the program and
functional managers
 Insures that a robust proposal gets
submitted to the REQUESTER on time
   Program/project costs involve both direct labor
   Each team member should understand overhead
rates
   If overhead rates are more than 50% of direct
regular time and not chargeable to overtime, then
overtime at 150% regular time may be cheaper
   Overhead rates in manufacturing can be 300-450%
(Indirect Costs)
   Building maintenance        Group insurance
   Building rent               Holidays
   Cafeteria                   Moving/storage exp.
   Clerical                    Personnel recruitment
   Consulting                  Retirement plans
   Corporate Salaries          Sick leave
   Depreciation of equip.      Telephone/Utilities
   Executive Salaries          Vacation
Why are Overhead Rates of Interest to
Project Management???
 These rates must be included in any project
cost calculation!!!
 The contractor is going to pay both your
yours is the winning bid
 Where do the costs associated with bidding
and proposing go? Does anybody pay for
them or are they just a SUNK cost???
Let’s REVIEW: What are the
Major Cost Components?
 Salary structure
 Labor hours required
 Cost of materials and support
Cost of Materials??
 Required Software
 Diet coke, pizza
Materials Support Costs
 Are submitted by month for each month of
the project
 An escalation factor for material costs must
be applied
Pricing out the Work—STEPS
(from Kerzner, p. 738)
 Provide a complete definition of the work
requirements
 Establish a logic network with checkpoints
 Develop the work breakdown structure
 Price out the WBS
Pricing out the Work--STEPS,
Cont’d
 Review WBS costs with each functional
manager
 Decide on the basic course of action
 Establish reasonable costs for each WBS
element
 Review the base case costs with upper-level
management
Pricing out the Work--STEPS,
Cont’d
 Negotiate with functional managers for
qualified personnel
 Develop the linear responsibility chart
 Develop the final detailed and PERT/CPM
schedules
 Establish pricing cost summary reports
 Document the result in a project plan
Smoothing out Department Man-
hours
   Ramp-up at project initiation and Ramp-down at
project completion cause step functions in
manpower requirements, as shown in Figure 14-8
   Functional managers attempt to SMOOTH this out
   QUESTION?? Does the department have
sufficient resources to fulfill the requirements?
Smoothing out Department Man-
hours
   ANOTHER QUESTION?? Can the
departments ramp-up fast enough?
The Pricing Review Procedure
 Based on Kerzner’s work
 Remember only 30 days to get the proposal
out and this is one of 13 steps
 Many contractors require the actual team
members to be identified in the proposal
 What solution comes to mind?
Systems Pricing
 The project pricing model (also called the
strategic planning model) acts as a
management information system
 Also provides management with an
invaluable tool for performing perturbation
analysis on the base case costs
Developing the
Supporting/Backup Costs
 Some proposals require backup support
 When required backup support must be
included in the pricing
 An issue is the type of contract
Types of contracts
 Fixed-price (developer assumes all of the
risk)
 Cost-plus (contractor pays for every hour
invested and thus assumes all the risk)
 An infinite array in between these
The Low-Bidder Dilemma
 The price of your contract will definitely
affect the viability of your proposal
 A low price on cost-plus type proposals is
suspect
 A low price on fixed-price contracts may be
perceived as impossible and undoable, or if
accepted will lead to a disaster
The Price on the Proposal is
always relative to:
 the competitive prices
 the customer budget
 the bidder’s cost estimate
 IN ANY CASE, LOW PRICING
WITHOUT MARKET INFORMATION IS
MEANINGLESS
If its a new market for the
Developer:
 Cost sharing may be an effective strategy
 Bidding below your actual costs is
commonplace
 Contractor’s objectives might include
system life cycle cost or unit production
cost
The Bottom Line on Price
   THE LOWEST BIDDER IS NOT
NECESSARILY THE AUTOMATIC WINNER
– Makes project a risky image regarding cost,
performance or schedule
   The ability to perform under contract is a definite
consideration
   A compliant, technically and managerially sound
proposal based on past experience, with realistic,
well-documented cost figures, is often chosen over
the lowest bidder
Special Problems
• Pricing must include an understanding of cost control--
how costs are billed back to the project
• There are three situations:
– Work is priced out at the department average, and all
work performed is charged to the project at the
department average, regardless of who performed the
work
– Work is priced out at the dept.. average, but all work
performed is billed back to the project at the actual
salary of the employees who performed the work
– Work is priced out at the actual salary of those
employees who will perform the work, and the cost is
billed back the same way.

» This is the ideal situation
This is as far as we will go with
these slides—ignore the
remainder
Estimating Pitfalls
• The “buy-in” decision is the most serious pitfall
because it means that the project will be under-
funded
• If the customer initially defines the requirements
and you (the developer) further refine them and
the customer doesn’t understand what you’ve
done, whose fault is it?
Estimating High-Risk Projects
• Validity of historical estimates determine the
difference between high-risk and low-risk
projects
• Estimating high-risk projects is commonly done
by means of the rolling wave or moving window
approach
– For a 12-month project the first six months are
estimated to level 5, while the last six months are
estimated to level two only.
– As the project proceeds more and more of the last six
months is estimated to level 5
– See Figure 14-13, Kerzner
Project Risks
   RISKS--Factors that increase the
probability that the project’s goals of time,
cost and performance will not be met
• See Figures 14-14, 14-15 and Table 14-13
(Especially useful)
Common Risks include:
 Poorly defined requirements
 Lack of qualified resources
 Lack of management support
 Poor estimating
 Inexperienced project manager
Tools to Aid in Risk
Identification
 Decision Support Systems
 Expected value measures
 Trend analysis/projection
 Independent reviews and audits
Six steps to risk management are:
 Identification of the risk
 Quantification of the risk
 Prioritizing the risk
 Developing a strategy for managing the risk
– A contingency plan
 Taking action
The Disaster of Applying the 10
% Solution to Project Estimates
   10% is taken from every on-going project to create
a budget out of thin air
   The result is havoc on top of chaos
   Most high-level executive committees do not
realize the impact of adopting the 10% solution
   A REDUCTION IN BUDGET MUST BE
ACCOMPANIED BY A TRADEOFF IN TIME
OR PERFORMANCE
The Disaster of the 10%
Solution, Cont’d
 90% of the budget generates 10% of the
desired service or performance levels and
the remaining 10% will generate the last
90% of the desired service or performance
 If there is FAT, i.e., padding, it may,
however, be possible to sustain a cut in the
project budget without major consequence
– Most projects do not have FAT
Cost vs. Performance
 I much prefer the word performance to
quality here
 A 10% reduction in cost can be expected to
produce much greater than a 10% reduction
in performance
More on the 10% Solution
 10% reduction solutions should be
undertaken only after a careful impact study
has been completed
 A far better choice is for the executive
committee to cancel or de-scope some
projects in order to release funds
14.19 Life-Cycle Costing (LCC)
 These are the total costs to the organization
for the ownership and acquisition of a
product over its full life
 Especially appropriate for in-house software
development projects, but is used in some
(outhouse) contracted projects as well
Life-cycle cost breakdown
 R & D Costs (Definition, Analysis)
 Production cost (Design)
 Construction cost (Programming and
testing)
 Operation and maintenance cost
 Product retirement cost
Life-cycle costing process
 Define the problem
 Define the requirements of the cost model
being used
 Collect historical data-cost relationships
 Develop estimates and test results
Successful applications of LCC
will:
 Provide downstream resource impact
visibility
 Provide life-cycle cost management
 Influence R&D decision making
 Support downstream strategic budgeting
Estimating Methods for LCC
   Method choice is based on the problem context,
operational considerations, etc.
   Informal estimating methods
–   Judgment based on experience
–   Analogy
–   SWAG method, ROM method
–   Rule-of-thumb method
   Formal estimating methods
– Detailed (from industrial engineering standards)
– Parametric
Figure 14-18
 For every \$12 DOD puts into R&D, \$28 are
needed downstream for production and \$60
for operation and support
 After Conceptual definition and
demonstration/validation, 85% of the
lifecycle decisions are made and cost
reduction opportunities are down to 22%
14.20 Logistics Support
   A frequent occurrence in software development
where the developer is paid to provide after-
market support in the form of operation and
maintenance on the product (deliverable)
   Recall that after the design phase 85% of the
deliverable’s life-cycle cost has been committed
and the majority of the total life-cycle cost is still
Performance metrics for Products
requiring Logistics Support
 Suppportability--the ability to maintain or
acquire the necessary human resources to
support the system
 Readiness--measure of how good we are at
keeping the product performing as planned
and how quickly we can make repairs
during a shutdown
Ten elements of logistics support
   Maintenance planning    Training and
   Manpower and             training support
personnel               Computer resources
support
   Supply support
 Facilities
   Support equipment       Packaging,
   Technical data           handling, storage
   See page 765             and transportation
Estimating --
   An iterative process
– Definition, Analysis, Design
 After Definition, 50-100% off
 After Analysis, 25-50% off
 After Medium level design--within 10%
 A good WBS is absolutely essential to do
estimating
Estimating Techniques
 Professional Judgment
 Developer estimate
 History (database)
 Formulas
Use of Professional Judgment
 Based on WBS, an expert judgment
estimate is made for each work package
 Amazingly accurate when experts are
available
 Often, however experts aren’t available
Developer Estimate
   The programmer assigned to the work
package will make every effort to complete
the task in the time he estimated it would
take
Use of History Database
 For this to work, your firm must keep a
history database
 The database should record how long each
 Break new projects up into tasks that have a
history database
 8 to 1 productivity ratio between best to
worst professional
Questions, Cont’d
   How much of the total time does Brooks devote to
Definition, Analysis and Design?
   1/3
   How much time to coding?
   1/6 to Coding
   How much time to testing?
   1/4 to component test and early system test
   1/4 to total system test
   So how much time are you going to devote to
Use of Formulas

   COCOMO--project cost, effort, schedule,
staffing for each of the phases:
–   Preliminary design
–   Detailed design
–   Code and unit test
–   System test
   COCOMO was developed by Barry Boehm
in 1981--COnstructive COst MOdel
Inputs to COCOMO
 Monthly cost of staff involved
 Factors indicating the general level of
complexity of the software
 Programming practices and tools used
 Experience of staff
 Lines of LOSC--rendering COCOMO
unusable
Function Points
 A user input
 User display
 Peripheral I/O
 Restructuring data
 Condition checking
 Calculation
 Branching
Function point approach--
BEFORE YOU LEAP
 Vendor is Gordon Group
 It must know how many LOSC are required
for each function point.
 It calculates LOSC based on function points
it knows about and feeds this into the
COCOMO algorithm
Estimacs from CA (Computer
Associates)
 Can take into account modern code
generation tools
 Determines effort, but also
–   Hardware required
–   financial break-even analysis
–   risk analysis
–   maintenance costs
   Expensive > \$20K
Estimating Programming:
Function Points
 D = C * ( G + J)
 D is the task duration in person-days
 C is the complexity of the task
 G is the assigned persons’ general
experience
 J is the assigned professional’s job
knowledge factor
Complexity
 Must break task down into its smallest
possible repeatable functions
 Then add up the complexity of each
function
 User input, user display, peripheral I/O,
restructuring data, condition checking,
calculation, etc.
   Repeatable functions are called function points.
   Function points are graded as SIMPLE, COMPLEX and VERY
COMPLEX
Productivity
 Your average programmer gets a
productivity factor of 1 for G
 Slower programmers get factors > 1
 Faster programmers get factors < 1
Formula method conclusions
 Will work if you develop accurate factors
 Can be used for any task from building a
house to developing software
 Depends on how well you granularize
Estimating The Analysis Phase
 Interviews
 Analyze Existing Documents and Systems
 Prepare Functional Specification
 Presentation
RATIOS

 PHASE                 PERCENTAGE
 Definition phase -- 10%
 Analysis phase -- 20%
 Design phase    -- 10%
 Programming     -- 20%
 System test     -- 17%
 Acceptance       -- 7%
 Operation        -- 16%
This breaks down to:
 PLAN -- 40%
 BUILD -- 20%
 TEST -- 40%
Another Rule of Thumb
 The time to design, document and code a
module =
 equals the time to debug it

   According to Gildersleeve
Can you use RATIOS for
Forecasting?
 Suppose you found that it took 20 days to
do definition.
 How long, based on ratios will it take to do
the project?
Estimating Rules
   Never use inexperienced persons to estimate
   Get group estimates if possible
   Never force an estimate on a programmer
   Never take an average of different estimates
   Granularize down to one week or less
   Always quote a range when giving estimates
Conclusions to Estimating
 Our weakest talent
 Estimating is iterative
 Estimating is still an art
Scheduling --
   Also assists with estimating, especially
when PM software is used
PM software supports
 WBS
 Gantt
 PERT
 Calendar(s)
 Resources and their assignments
PERT
   Use activity on node approach
– doesn’t require dummy activities
 Understand what float is--it is slack
 Critical path is the longest path
 shows precedent activities, relationships
 doesn’t show what will be done when, by
whom
Resource allocation
 Assign tasks to individuals whose skill level
 Assign similar tasks to the same person
reliable people
 Assign tasks that communicate to the same
individual to minimize people’s interaction
 Don’t assign too many different tasks to any
one individual
manpower
member on a professional team
 If it takes 10 person days for one person, it
will take 12 person days for two people,
14.4 person days for three people, etc.
 More resources, gets the project done
sooner, USUALLY
 But it also costs more
 The PM must come up with the best
balance, depending on the priorities set by
management or the user
Shortening the duration of
projects
 Fast tracking
 Crashing
– Adding resources to the critical path
– Allowing your current CP teams members to
work overtime
Crashing projects
 Crash tasks on the critical path only, only as
long as no other path becomes critical
 If other paths become critical, the analyst
must crash those as well
Gantt Chart
 a time bar chart
 Invented by Henry Gantt in 1910
 Determine the units of time
 Mark all known calendar events at bottom
 Schedule each activity from PERT
Use three sets of Gantt’s
 one for yourself alone, with all float and
contingency visible
 second for the individuals involved--their
resource Gantt, contingencies hidden
 third for distribution to upper management--
contingencies hidden
 Include a 10% contingency into all
estimates
Conclusions to Scheduling
 Use PM software--worth every penny
 Do at least one PERT and one GANTT
manually, just to get a feel for the process
Recitation
   What is the probability of completing a project by
its estimated completion time??
   What is the formula for calculating the completion
time for a PERT network?
   What is the formula for calculating the standard
deviation of the completion time for a PERT
network?
   Name some processes that are part of project
integration management

```
To top